Daily Papers

With the recent success of the pre-training technique for NLP and image-linguistic tasks, some video-linguistic pre-training works are gradually developed to improve video-text related downstream tasks. However, most of the existing multimodal models are pre-trained for understanding tasks, leading to a pretrain-finetune discrepancy for generation tasks. This paper proposes UniVL: a Unified Video and Language pre-training model for both multimodal understanding and generation. It comprises four components, including two single-modal encoders, a cross encoder, and a decoder with the Transformer backbone. Five objectives, including video-text joint, conditioned masked language model (CMLM), conditioned masked frame model (CMFM), video-text alignment, and language reconstruction, are designed to train each of the components. We further develop two pre-training strategies, stage by stage pre-training (StagedP) and enhanced video representation (EnhancedV), to make the training process of the UniVL more effective. The pre-train is carried out on a sizeable instructional video dataset HowTo100M. Experimental results demonstrate that the UniVL can learn strong video-text representation and achieves state-of-the-art results on five downstream tasks.

We introduce MarDini, a new family of video diffusion models that integrate the advantages of masked auto-regression (MAR) into a unified diffusion model (DM) framework. Here, MAR handles temporal planning, while DM focuses on spatial generation in an asymmetric network design: i) a MAR-based planning model containing most of the parameters generates planning signals for each masked frame using low-resolution input; ii) a lightweight generation model uses these signals to produce high-resolution frames via diffusion de-noising. MarDini's MAR enables video generation conditioned on any number of masked frames at any frame positions: a single model can handle video interpolation (e.g., masking middle frames), image-to-video generation (e.g., masking from the second frame onward), and video expansion (e.g., masking half the frames). The efficient design allocates most of the computational resources to the low-resolution planning model, making computationally expensive but important spatio-temporal attention feasible at scale. MarDini sets a new state-of-the-art for video interpolation; meanwhile, within few inference steps, it efficiently generates videos on par with those of much more expensive advanced image-to-video models.

Recently, masked video modeling has been widely explored and significantly improved the model's understanding ability of visual regions at a local level. However, existing methods usually adopt random masking and follow the same reconstruction paradigm to complete the masked regions, which do not leverage the correlations between cross-modal content. In this paper, we present Mask for Semantics Completion (MASCOT) based on semantic-based masked modeling. Specifically, after applying attention-based video masking to generate high-informed and low-informed masks, we propose Informed Semantics Completion to recover masked semantics information. The recovery mechanism is achieved by aligning the masked content with the unmasked visual regions and corresponding textual context, which makes the model capture more text-related details at a patch level. Additionally, we shift the emphasis of reconstruction from irrelevant backgrounds to discriminative parts to ignore regions with low-informed masks. Furthermore, we design dual-mask co-learning to incorporate video cues under different masks and learn more aligned video representation. Our MASCOT performs state-of-the-art performance on four major text-video retrieval benchmarks, including MSR-VTT, LSMDC, ActivityNet, and DiDeMo. Extensive ablation studies demonstrate the effectiveness of the proposed schemes.

Masked Video Autoencoder (MVA) approaches have demonstrated their potential by significantly outperforming previous video representation learning methods. However, they waste an excessive amount of computations and memory in predicting uninformative tokens/frames due to random masking strategies. (e.g., over 16 nodes with 128 NVIDIA A100 GPUs). To resolve this issue, we exploit the unequal information density among the patches in videos and propose EVEREST, a surprisingly efficient MVA approach for video representation learning that finds tokens containing rich motion features and discards uninformative ones during both pre-training and fine-tuning. We further present an information-intensive frame selection strategy that allows the model to focus on informative and causal frames with minimal redundancy. Our method significantly reduces the computation and memory requirements of MVA, enabling the pre-training and fine-tuning on a single machine with 8 GPUs while achieving comparable performance to computation- and memory-heavy baselines on multiple benchmarks and the uncurated Ego4D dataset. We hope that our work contributes to reducing the barrier to further research on video understanding.

We present Masked Feature Prediction (MaskFeat) for self-supervised pre-training of video models. Our approach first randomly masks out a portion of the input sequence and then predicts the feature of the masked regions. We study five different types of features and find Histograms of Oriented Gradients (HOG), a hand-crafted feature descriptor, works particularly well in terms of both performance and efficiency. We observe that the local contrast normalization in HOG is essential for good results, which is in line with earlier work using HOG for visual recognition. Our approach can learn abundant visual knowledge and drive large-scale Transformer-based models. Without using extra model weights or supervision, MaskFeat pre-trained on unlabeled videos achieves unprecedented results of 86.7% with MViT-L on Kinetics-400, 88.3% on Kinetics-600, 80.4% on Kinetics-700, 39.8 mAP on AVA, and 75.0% on SSv2. MaskFeat further generalizes to image input, which can be interpreted as a video with a single frame and obtains competitive results on ImageNet.

Self-supervised learning of image representations by predicting future frames is a promising direction but still remains a challenge. This is because of the under-determined nature of frame prediction; multiple potential futures can arise from a single current frame. To tackle this challenge, in this paper, we revisit the idea of stochastic video generation that learns to capture uncertainty in frame prediction and explore its effectiveness for representation learning. Specifically, we design a framework that trains a stochastic frame prediction model to learn temporal information between frames. Moreover, to learn dense information within each frame, we introduce an auxiliary masked image modeling objective along with a shared decoder architecture. We find this architecture allows for combining both objectives in a synergistic and compute-efficient manner. We demonstrate the effectiveness of our framework on a variety of tasks from video label propagation and vision-based robot learning domains, such as video segmentation, pose tracking, vision-based robotic locomotion, and manipulation tasks. Code is available on the project webpage: https://sites.google.com/view/2024rsp.

Self-attention is of vital importance in semantic segmentation as it enables modeling of long-range context, which translates into improved performance. We argue that it is equally important to model short-range context, especially to tackle cases where not only the regions of interest are small and ambiguous, but also when there exists an imbalance between the semantic classes. To this end, we propose Masked Supervised Learning (MaskSup), an effective single-stage learning paradigm that models both short- and long-range context, capturing the contextual relationships between pixels via random masking. Experimental results demonstrate the competitive performance of MaskSup against strong baselines in both binary and multi-class segmentation tasks on three standard benchmark datasets, particularly at handling ambiguous regions and retaining better segmentation of minority classes with no added inference cost. In addition to segmenting target regions even when large portions of the input are masked, MaskSup is also generic and can be easily integrated into a variety of semantic segmentation methods. We also show that the proposed method is computationally efficient, yielding an improved performance by 10\% on the mean intersection-over-union (mIoU) while requiring 3times less learnable parameters.

Masked modeling has emerged as a powerful self-supervised learning framework, but existing methods largely rely on random masking, disregarding the structural properties of different modalities. In this work, we introduce structured noise-based masking, a simple yet effective approach that naturally aligns with the spatial, temporal, and spectral characteristics of video and audio data. By filtering white noise into distinct color noise distributions, we generate structured masks that preserve modality-specific patterns without requiring handcrafted heuristics or access to the data. Our approach improves the performance of masked video and audio modeling frameworks without any computational overhead. Extensive experiments demonstrate that structured noise masking achieves consistent improvement over random masking for standard and advanced masked modeling methods, highlighting the importance of modality-aware masking strategies for representation learning.

How to learn discriminative video representation from unlabeled videos is challenging but crucial for video analysis. The latest attempts seek to learn a representation model by predicting the appearance contents in the masked regions. However, simply masking and recovering appearance contents may not be sufficient to model temporal clues as the appearance contents can be easily reconstructed from a single frame. To overcome this limitation, we present Masked Motion Encoding (MME), a new pre-training paradigm that reconstructs both appearance and motion information to explore temporal clues. In MME, we focus on addressing two critical challenges to improve the representation performance: 1) how to well represent the possible long-term motion across multiple frames; and 2) how to obtain fine-grained temporal clues from sparsely sampled videos. Motivated by the fact that human is able to recognize an action by tracking objects' position changes and shape changes, we propose to reconstruct a motion trajectory that represents these two kinds of change in the masked regions. Besides, given the sparse video input, we enforce the model to reconstruct dense motion trajectories in both spatial and temporal dimensions. Pre-trained with our MME paradigm, the model is able to anticipate long-term and fine-grained motion details. Code is available at https://github.com/XinyuSun/MME.

Recognizing multiple objects in an image is challenging due to occlusions, and becomes even more so when the objects are small. While promising, existing multi-label image recognition models do not explicitly learn context-based representations, and hence struggle to correctly recognize small and occluded objects. Intuitively, recognizing occluded objects requires knowledge of partial input, and hence context. Motivated by this intuition, we propose Masked Supervised Learning (MSL), a single-stage, model-agnostic learning paradigm for multi-label image recognition. The key idea is to learn context-based representations using a masked branch and to model label co-occurrence using label consistency. Experimental results demonstrate the simplicity, applicability and more importantly the competitive performance of MSL against previous state-of-the-art methods on standard multi-label image recognition benchmarks. In addition, we show that MSL is robust to random masking and demonstrate its effectiveness in recognizing non-masked objects. Code and pretrained models are available on GitHub.

Temporal action segmentation is crucial for understanding long-form videos. Previous works on this task commonly adopt an iterative refinement paradigm by using multi-stage models. We propose a novel framework via denoising diffusion models, which nonetheless shares the same inherent spirit of such iterative refinement. In this framework, action predictions are iteratively generated from random noise with input video features as conditions. To enhance the modeling of three striking characteristics of human actions, including the position prior, the boundary ambiguity, and the relational dependency, we devise a unified masking strategy for the conditioning inputs in our framework. Extensive experiments on three benchmark datasets, i.e., GTEA, 50Salads, and Breakfast, are performed and the proposed method achieves superior or comparable results to state-of-the-art methods, showing the effectiveness of a generative approach for action segmentation.

Recent video masked autoencoder (MAE) works have designed improved masking algorithms focused on saliency. These works leverage visual cues such as motion to mask the most salient regions. However, the robustness of such visual cues depends on how often input videos match underlying assumptions. On the other hand, natural language description is an information dense representation of video that implicitly captures saliency without requiring modality-specific assumptions, and has not been explored yet for video MAE. To this end, we introduce a novel text-guided masking algorithm (TGM) that masks the video regions with highest correspondence to paired captions. Without leveraging any explicit visual cues for saliency, our TGM is competitive with state-of-the-art masking algorithms such as motion-guided masking. To further benefit from the semantics of natural language for masked reconstruction, we next introduce a unified framework for joint MAE and masked video-text contrastive learning. We show that across existing masking algorithms, unifying MAE and masked video-text contrastive learning improves downstream performance compared to pure MAE on a variety of video recognition tasks, especially for linear probe. Within this unified framework, our TGM achieves the best relative performance on five action recognition and one egocentric datasets, highlighting the complementary nature of natural language for masked video modeling.

The ability to predict future visual observations conditioned on past observations and motor commands can enable embodied agents to plan solutions to a variety of tasks in complex environments. This work shows that we can create good video prediction models by pre-training transformers via masked visual modeling. Our approach, named MaskViT, is based on two simple design decisions. First, for memory and training efficiency, we use two types of window attention: spatial and spatiotemporal. Second, during training, we mask a variable percentage of tokens instead of a fixed mask ratio. For inference, MaskViT generates all tokens via iterative refinement where we incrementally decrease the masking ratio following a mask scheduling function. On several datasets we demonstrate that MaskViT outperforms prior works in video prediction, is parameter efficient, and can generate high-resolution videos (256x256). Further, we demonstrate the benefits of inference speedup (up to 512x) due to iterative decoding by using MaskViT for planning on a real robot. Our work suggests that we can endow embodied agents with powerful predictive models by leveraging the general framework of masked visual modeling with minimal domain knowledge.

We present an efficient method of utilizing pretrained language models, where we learn selective binary masks for pretrained weights in lieu of modifying them through finetuning. Extensive evaluations of masking BERT and RoBERTa on a series of NLP tasks show that our masking scheme yields performance comparable to finetuning, yet has a much smaller memory footprint when several tasks need to be inferred simultaneously. Through intrinsic evaluations, we show that representations computed by masked language models encode information necessary for solving downstream tasks. Analyzing the loss landscape, we show that masking and finetuning produce models that reside in minima that can be connected by a line segment with nearly constant test accuracy. This confirms that masking can be utilized as an efficient alternative to finetuning.

Pretraining general-purpose visual features has become a crucial part of tackling many computer vision tasks. While one can learn such features on the extensively-annotated ImageNet dataset, recent approaches have looked at ways to allow for noisy, fewer, or even no annotations to perform such pretraining. Starting from the observation that captioned images are easily crawlable, we argue that this overlooked source of information can be exploited to supervise the training of visual representations. To do so, motivated by the recent progresses in language models, we introduce {\em image-conditioned masked language modeling} (ICMLM) -- a proxy task to learn visual representations over image-caption pairs. ICMLM consists in predicting masked words in captions by relying on visual cues. To tackle this task, we propose hybrid models, with dedicated visual and textual encoders, and we show that the visual representations learned as a by-product of solving this task transfer well to a variety of target tasks. Our experiments confirm that image captions can be leveraged to inject global and localized semantic information into visual representations. Project website: https://europe.naverlabs.com/icmlm.

In the domain of computer vision, the restoration of missing information in video frames is a critical challenge, particularly in applications such as autonomous driving and surveillance systems. This paper introduces the Siamese Masked Conditional Variational Autoencoder (SiamMCVAE), leveraging a siamese architecture with twin encoders based on vision transformers. This innovative design enhances the model's ability to comprehend lost content by capturing intrinsic similarities between paired frames. SiamMCVAE proficiently reconstructs missing elements in masked frames, effectively addressing issues arising from camera malfunctions through variational inferences. Experimental results robustly demonstrate the model's effectiveness in restoring missing information, thus enhancing the resilience of computer vision systems. The incorporation of Siamese Vision Transformer (SiamViT) encoders in SiamMCVAE exemplifies promising potential for addressing real-world challenges in computer vision, reinforcing the adaptability of autonomous systems in dynamic environments.

Recently, Vision Large Language Models (VLLMs) integrated with vision encoders have shown promising performance in vision understanding. The key of VLLMs is to encode visual content into sequences of visual tokens, enabling VLLMs to simultaneously process both visual and textual content. However, understanding videos, especially long videos, remain a challenge to VLLMs as the number of visual tokens grows rapidly when encoding videos, resulting in the risk of exceeding the context window of VLLMs and introducing heavy computation burden. To restrict the number of visual tokens, existing VLLMs either: (1) uniformly downsample videos into a fixed number of frames or (2) reducing the number of visual tokens encoded from each frame. We argue the former solution neglects the rich temporal cue in videos and the later overlooks the spatial details in each frame. In this work, we present Balanced-VLLM (B-VLLM): a novel VLLM framework that aims to effectively leverage task relevant spatio-temporal cues while restricting the number of visual tokens under the VLLM context window length. At the core of our method, we devise a text-conditioned adaptive frame selection module to identify frames relevant to the visual understanding task. The selected frames are then de-duplicated using a temporal frame token merging technique. The visual tokens of the selected frames are processed through a spatial token sampling module and an optional spatial token merging strategy to achieve precise control over the token count. Experimental results show that B-VLLM is effective in balancing the number of frames and visual tokens in video understanding, yielding superior performance on various video understanding benchmarks. Our code is available at https://github.com/zhuqiangLu/B-VLLM.

Generating controllable character animation from a reference image and motion guidance remains a challenging task due to the inherent difficulty of injecting appearance and motion cues into video diffusion models. Prior works often rely on complex architectures, explicit guider modules, or multi-stage processing pipelines, which increase structural overhead and hinder deployment. Inspired by the strong visual context modeling capacity of pre-trained video diffusion transformers, we propose FramePrompt, a minimalist yet powerful framework that treats reference images, skeleton-guided motion, and target video clips as a unified visual sequence. By reformulating animation as a conditional future prediction task, we bypass the need for guider networks and structural modifications. Experiments demonstrate that our method significantly outperforms representative baselines across various evaluation metrics while also simplifying training. Our findings highlight the effectiveness of sequence-level visual conditioning and demonstrate the potential of pre-trained models for controllable animation without architectural changes.

Recent advances in generative AI have significantly enhanced image and video editing, particularly in the context of text prompt control. State-of-the-art approaches predominantly rely on diffusion models to accomplish these tasks. However, the computational demands of diffusion-based methods are substantial, often necessitating large-scale paired datasets for training, and therefore challenging the deployment in practical applications. This study addresses this challenge by breaking down the text-based video editing process into two separate stages. In the first stage, we leverage an existing text-to-image diffusion model to simultaneously edit a few keyframes without additional fine-tuning. In the second stage, we introduce an efficient model called MaskINT, which is built on non-autoregressive masked generative transformers and specializes in frame interpolation between the keyframes, benefiting from structural guidance provided by intermediate frames. Our comprehensive set of experiments illustrates the efficacy and efficiency of MaskINT when compared to other diffusion-based methodologies. This research offers a practical solution for text-based video editing and showcases the potential of non-autoregressive masked generative transformers in this domain.

The objective of this paper is to be able to separate a video into its natural layers, and to control which of the separated layers to attend to. For example, to be able to separate reflections, transparency or object motion. We make the following three contributions: (i) we introduce a new structured neural network architecture that explicitly incorporates layers (as spatial masks) into its design. This improves separation performance over previous general purpose networks for this task; (ii) we demonstrate that we can augment the architecture to leverage external cues such as audio for controllability and to help disambiguation; and (iii) we experimentally demonstrate the effectiveness of our approach and training procedure with controlled experiments while also showing that the proposed model can be successfully applied to real-word applications such as reflection removal and action recognition in cluttered scenes.

Human motion prediction is a classical problem in computer vision and computer graphics, which has a wide range of practical applications. Previous effects achieve great empirical performance based on an encoding-decoding style. The methods of this style work by first encoding previous motions to latent representations and then decoding the latent representations into predicted motions. However, in practice, they are still unsatisfactory due to several issues, including complicated loss constraints, cumbersome training processes, and scarce switch of different categories of motions in prediction. In this paper, to address the above issues, we jump out of the foregoing style and propose a novel framework from a new perspective. Specifically, our framework works in a masked completion fashion. In the training stage, we learn a motion diffusion model that generates motions from random noise. In the inference stage, with a denoising procedure, we make motion prediction conditioning on observed motions to output more continuous and controllable predictions. The proposed framework enjoys promising algorithmic properties, which only needs one loss in optimization and is trained in an end-to-end manner. Additionally, it accomplishes the switch of different categories of motions effectively, which is significant in realistic tasks, e.g., the animation task. Comprehensive experiments on benchmarks confirm the superiority of the proposed framework. The project page is available at https://lhchen.top/Human-MAC.

For a given scene, humans can easily reason for the locations and pose to place objects. Designing a computational model to reason about these affordances poses a significant challenge, mirroring the intuitive reasoning abilities of humans. This work tackles the problem of realistic human insertion in a given background scene termed as Semantic Human Placement. This task is extremely challenging given the diverse backgrounds, scale, and pose of the generated person and, finally, the identity preservation of the person. We divide the problem into the following two stages i) learning semantic masks using text guidance for localizing regions in the image to place humans and ii) subject-conditioned inpainting to place a given subject adhering to the scene affordance within the semantic masks. For learning semantic masks, we leverage rich object-scene priors learned from the text-to-image generative models and optimize a novel parameterization of the semantic mask, eliminating the need for large-scale training. To the best of our knowledge, we are the first ones to provide an effective solution for realistic human placements in diverse real-world scenes. The proposed method can generate highly realistic scene compositions while preserving the background and subject identity. Further, we present results for several downstream tasks - scene hallucination from a single or multiple generated persons and text-based attribute editing. With extensive comparisons against strong baselines, we show the superiority of our method in realistic human placement.

Several video understanding tasks, such as natural language temporal video grounding, temporal activity localization, and audio description generation, require "temporally dense" reasoning over frames sampled at high temporal resolution. However, computing frame-level features for these tasks is computationally expensive given the temporal resolution requirements. In this paper, we make three contributions to reduce the cost of computing features for temporally dense tasks. First, we introduce a vision transformer (ViT) architecture, dubbed ResidualViT, that leverages the large temporal redundancy in videos to efficiently compute temporally dense frame-level features. Our architecture incorporates (i) learnable residual connections that ensure temporal consistency across consecutive frames and (ii) a token reduction module that enhances processing speed by selectively discarding temporally redundant information while reusing weights of a pretrained foundation model. Second, we propose a lightweight distillation strategy to approximate the frame-level features of the original foundation model. Finally, we evaluate our approach across four tasks and five datasets, in both zero-shot and fully supervised settings, demonstrating significant reductions in computational cost (up to 60%) and improvements in inference speed (up to 2.5x faster), all while closely approximating the accuracy of the original foundation model.

Video prediction is a challenging task. The quality of video frames from current state-of-the-art (SOTA) generative models tends to be poor and generalization beyond the training data is difficult. Furthermore, existing prediction frameworks are typically not capable of simultaneously handling other video-related tasks such as unconditional generation or interpolation. In this work, we devise a general-purpose framework called Masked Conditional Video Diffusion (MCVD) for all of these video synthesis tasks using a probabilistic conditional score-based denoising diffusion model, conditioned on past and/or future frames. We train the model in a manner where we randomly and independently mask all the past frames or all the future frames. This novel but straightforward setup allows us to train a single model that is capable of executing a broad range of video tasks, specifically: future/past prediction -- when only future/past frames are masked; unconditional generation -- when both past and future frames are masked; and interpolation -- when neither past nor future frames are masked. Our experiments show that this approach can generate high-quality frames for diverse types of videos. Our MCVD models are built from simple non-recurrent 2D-convolutional architectures, conditioning on blocks of frames and generating blocks of frames. We generate videos of arbitrary lengths autoregressively in a block-wise manner. Our approach yields SOTA results across standard video prediction and interpolation benchmarks, with computation times for training models measured in 1-12 days using le 4 GPUs. Project page: https://mask-cond-video-diffusion.github.io ; Code : https://github.com/voletiv/mcvd-pytorch

Masked Language Models (MLMs) have achieved remarkable success in many self-supervised representation learning tasks. MLMs are trained by randomly masking portions of the input sequences with [MASK] tokens and learning to reconstruct the original content based on the remaining context. This paper explores the impact of [MASK] tokens on MLMs. Analytical studies show that masking tokens can introduce the corrupted semantics problem, wherein the corrupted context may convey multiple, ambiguous meanings. This problem is also a key factor affecting the performance of MLMs on downstream tasks. Based on these findings, we propose a novel enhanced-context MLM, ExLM. Our approach expands [MASK] tokens in the input context and models the dependencies between these expanded states. This enhancement increases context capacity and enables the model to capture richer semantic information, effectively mitigating the corrupted semantics problem during pre-training. Experimental results demonstrate that ExLM achieves significant performance improvements in both text modeling and SMILES modeling tasks. Further analysis confirms that ExLM enriches semantic representations through context enhancement, and effectively reduces the semantic multimodality commonly observed in MLMs.

In generative models, two paradigms have gained attraction in various applications: next-set prediction-based Masked Generative Models and next-noise prediction-based Non-Autoregressive Models, e.g., Diffusion Models. In this work, we propose using discrete-state models to connect them and explore their scalability in the vision domain. First, we conduct a step-by-step analysis in a unified design space across two types of models including timestep-independence, noise schedule, temperature, guidance strength, etc in a scalable manner. Second, we re-cast typical discriminative tasks, e.g., image segmentation, as an unmasking process from [MASK]tokens on a discrete-state model. This enables us to perform various sampling processes, including flexible conditional sampling by only training once to model the joint distribution. All aforementioned explorations lead to our framework named Discrete Interpolants, which enables us to achieve state-of-the-art or competitive performance compared to previous discrete-state based methods in various benchmarks, like ImageNet256, MS COCO, and video dataset FaceForensics. In summary, by leveraging [MASK] in discrete-state models, we can bridge Masked Generative and Non-autoregressive Diffusion models, as well as generative and discriminative tasks.

We present Masked Frequency Modeling (MFM), a unified frequency-domain-based approach for self-supervised pre-training of visual models. Instead of randomly inserting mask tokens to the input embeddings in the spatial domain, in this paper, we shift the perspective to the frequency domain. Specifically, MFM first masks out a portion of frequency components of the input image and then predicts the missing frequencies on the frequency spectrum. Our key insight is that predicting masked components in the frequency domain is more ideal to reveal underlying image patterns rather than predicting masked patches in the spatial domain, due to the heavy spatial redundancy. Our findings suggest that with the right configuration of mask-and-predict strategy, both the structural information within high-frequency components and the low-level statistics among low-frequency counterparts are useful in learning good representations. For the first time, MFM demonstrates that, for both ViT and CNN, a simple non-Siamese framework can learn meaningful representations even using none of the following: (i) extra data, (ii) extra model, (iii) mask token. Experimental results on image classification and semantic segmentation, as well as several robustness benchmarks show the competitive performance and advanced robustness of MFM compared with recent masked image modeling approaches. Furthermore, we also comprehensively investigate the effectiveness of classical image restoration tasks for representation learning from a unified frequency perspective and reveal their intriguing relations with our MFM approach.

Current video-based Masked Autoencoders (MAEs) primarily focus on learning effective spatiotemporal representations from a visual perspective, which may lead the model to prioritize general spatial-temporal patterns but often overlook nuanced semantic attributes like specific interactions or sequences that define actions - such as action-specific features that align more closely with human cognition for space-time correspondence. This can limit the model's ability to capture the essence of certain actions that are contextually rich and continuous. Humans are capable of mapping visual concepts, object view invariance, and semantic attributes available in static instances to comprehend natural dynamic scenes or videos. Existing MAEs for videos and static images rely on separate datasets for videos and images, which may lack the rich semantic attributes necessary for fully understanding the learned concepts, especially when compared to using video and corresponding sampled frame images together. To this end, we propose CrossVideoMAE an end-to-end self-supervised cross-modal contrastive learning MAE that effectively learns both video-level and frame-level rich spatiotemporal representations and semantic attributes. Our method integrates mutual spatiotemporal information from videos with spatial information from sampled frames within a feature-invariant space, while encouraging invariance to augmentations within the video domain. This objective is achieved through jointly embedding features of visible tokens and combining feature correspondence within and across modalities, which is critical for acquiring rich, label-free guiding signals from both video and frame image modalities in a self-supervised manner. Extensive experiments demonstrate that our approach surpasses previous state-of-the-art methods and ablation studies validate the effectiveness of our approach.

A key challenge with procedure planning in instructional videos lies in how to handle a large decision space consisting of a multitude of action types that belong to various tasks. To understand real-world video content, an AI agent must proficiently discern these action types (e.g., pour milk, pour water, open lid, close lid, etc.) based on brief visual observation. Moreover, it must adeptly capture the intricate semantic relation of the action types and task goals, along with the variable action sequences. Recently, notable progress has been made via the integration of diffusion models and visual representation learning to address the challenge. However, existing models employ rudimentary mechanisms to utilize task information to manage the decision space. To overcome this limitation, we introduce a simple yet effective enhancement - a masked diffusion model. The introduced mask acts akin to a task-oriented attention filter, enabling the diffusion/denoising process to concentrate on a subset of action types. Furthermore, to bolster the accuracy of task classification, we harness more potent visual representation learning techniques. In particular, we learn a joint visual-text embedding, where a text embedding is generated by prompting a pre-trained vision-language model to focus on human actions. We evaluate the method on three public datasets and achieve state-of-the-art performance on multiple metrics. Code is available at https://github.com/ffzzy840304/Masked-PDPP.

Video prediction aims to generate realistic future frames by learning dynamic visual patterns. One fundamental challenge is to deal with future uncertainty: How should a model behave when there are multiple correct, equally probable future? We propose an Appearance-Motion Conditional GAN to address this challenge. We provide appearance and motion information as conditions that specify how the future may look like, reducing the level of uncertainty. Our model consists of a generator, two discriminators taking charge of appearance and motion pathways, and a perceptual ranking module that encourages videos of similar conditions to look similar. To train our model, we develop a novel conditioning scheme that consists of different combinations of appearance and motion conditions. We evaluate our model using facial expression and human action datasets and report favorable results compared to existing methods.

Multimodal Large Language Models (MLLMs) have demonstrated significant success in visual understanding tasks. However, challenges persist in adapting these models for video comprehension due to the large volume of data and temporal complexity. Existing Video-LLMs using uniform frame sampling often struggle to capture the query-related crucial spatiotemporal clues of videos effectively. In this paper, we introduce Q-Frame, a novel approach for adaptive frame selection and multi-resolution scaling tailored to the video's content and the specific query. Q-Frame employs a training-free, plug-and-play strategy generated by a text-image matching network like CLIP, utilizing the Gumbel-Max trick for efficient frame selection. Q-Frame allows Video-LLMs to process more frames without exceeding computational limits, thereby preserving critical temporal and spatial information. We demonstrate Q-Frame's effectiveness through extensive experiments on benchmark datasets, including MLVU, LongVideoBench, and Video-MME, illustrating its superiority over existing methods and its applicability across various video understanding tasks.

The joint-embedding predictive architecture (JEPA) recently has shown impressive results in extracting visual representations from unlabeled imagery under a masking strategy. However, we reveal its disadvantages, notably its insufficient understanding of local semantics. This deficiency originates from masked modeling in the embedding space, resulting in a reduction of discriminative power and can even lead to the neglect of critical local semantics. To bridge this gap, we introduce DMT-JEPA, a novel masked modeling objective rooted in JEPA, specifically designed to generate discriminative latent targets from neighboring information. Our key idea is simple: we consider a set of semantically similar neighboring patches as a target of a masked patch. To be specific, the proposed DMT-JEPA (a) computes feature similarities between each masked patch and its corresponding neighboring patches to select patches having semantically meaningful relations, and (b) employs lightweight cross-attention heads to aggregate features of neighboring patches as the masked targets. Consequently, DMT-JEPA demonstrates strong discriminative power, offering benefits across a diverse spectrum of downstream tasks. Through extensive experiments, we demonstrate our effectiveness across various visual benchmarks, including ImageNet-1K image classification, ADE20K semantic segmentation, and COCO object detection tasks. Code is available at: https://github.com/DMTJEPA/DMTJEPA.

Joint image-text embedding is the bedrock for most Vision-and-Language (V+L) tasks, where multimodality inputs are simultaneously processed for joint visual and textual understanding. In this paper, we introduce UNITER, a UNiversal Image-TExt Representation, learned through large-scale pre-training over four image-text datasets (COCO, Visual Genome, Conceptual Captions, and SBU Captions), which can power heterogeneous downstream V+L tasks with joint multimodal embeddings. We design four pre-training tasks: Masked Language Modeling (MLM), Masked Region Modeling (MRM, with three variants), Image-Text Matching (ITM), and Word-Region Alignment (WRA). Different from previous work that applies joint random masking to both modalities, we use conditional masking on pre-training tasks (i.e., masked language/region modeling is conditioned on full observation of image/text). In addition to ITM for global image-text alignment, we also propose WRA via the use of Optimal Transport (OT) to explicitly encourage fine-grained alignment between words and image regions during pre-training. Comprehensive analysis shows that both conditional masking and OT-based WRA contribute to better pre-training. We also conduct a thorough ablation study to find an optimal combination of pre-training tasks. Extensive experiments show that UNITER achieves new state of the art across six V+L tasks (over nine datasets), including Visual Question Answering, Image-Text Retrieval, Referring Expression Comprehension, Visual Commonsense Reasoning, Visual Entailment, and NLVR^2. Code is available at https://github.com/ChenRocks/UNITER.

The canonical approach to video captioning dictates a caption generation model to learn from offline-extracted dense video features. These feature extractors usually operate on video frames sampled at a fixed frame rate and are often trained on image/video understanding tasks, without adaption to video captioning data. In this work, we present SwinBERT, an end-to-end transformer-based model for video captioning, which takes video frame patches directly as inputs, and outputs a natural language description. Instead of leveraging multiple 2D/3D feature extractors, our method adopts a video transformer to encode spatial-temporal representations that can adapt to variable lengths of video input without dedicated design for different frame rates. Based on this model architecture, we show that video captioning can benefit significantly from more densely sampled video frames as opposed to previous successes with sparsely sampled video frames for video-and-language understanding tasks (e.g., video question answering). Moreover, to avoid the inherent redundancy in consecutive video frames, we propose adaptively learning a sparse attention mask and optimizing it for task-specific performance improvement through better long-range video sequence modeling. Through extensive experiments on 5 video captioning datasets, we show that SwinBERT achieves across-the-board performance improvements over previous methods, often by a large margin. The learned sparse attention masks in addition push the limit to new state of the arts, and can be transferred between different video lengths and between different datasets. Code is available at https://github.com/microsoft/SwinBERT

Video-language pre-training is crucial for learning powerful multi-modal representation. However, it typically requires a massive amount of computation. In this paper, we develop SMAUG, an efficient pre-training framework for video-language models. The foundation component in SMAUG is masked autoencoders. Different from prior works which only mask textual inputs, our masking strategy considers both visual and textual modalities, providing a better cross-modal alignment and saving more pre-training costs. On top of that, we introduce a space-time token sparsification module, which leverages context information to further select only "important" spatial regions and temporal frames for pre-training. Coupling all these designs allows our method to enjoy both competitive performances on text-to-video retrieval and video question answering tasks, and much less pre-training costs by 1.9X or more. For example, our SMAUG only needs about 50 NVIDIA A6000 GPU hours for pre-training to attain competitive performances on these two video-language tasks across six popular benchmarks.

Recent progress in Large Multi-modal Models (LMMs) has enabled effective vision-language reasoning, yet the ability to understand video content remains constrained by suboptimal frame selection strategies. Existing approaches often rely on static heuristics or external retrieval modules to feed frame information into video-LLMs, which may fail to provide the query-relevant information. In this work, we introduce ReFoCUS (Reinforcement-guided Frame Optimization for Contextual UnderStanding), a novel frame-level policy optimization framework that shifts the optimization target from textual responses to visual input selection. ReFoCUS learns a frame selection policy via reinforcement learning, using reward signals derived from a reference LMM to reflect the model's intrinsic preferences for frames that best support temporally grounded responses. To efficiently explore the large combinatorial frame space, we employ an autoregressive, conditional selection architecture that ensures temporal coherence while reducing complexity. Our approach does not require explicit supervision at the frame-level and consistently improves reasoning performance across multiple video QA benchmarks, highlighting the benefits of aligning frame selection with model-internal utility.

We introduce an attention-based method that uses learned binary attention masks to ensure that only attended image regions influence the prediction. Context can strongly affect object perception, sometimes leading to biased representations, particularly when objects appear in out-of-distribution backgrounds. At the same time, many image-level object-centric tasks require identifying relevant regions, often requiring context. To address this conundrum, we propose a two-stage framework: stage 1 processes the full image to discover object parts and identify task-relevant regions, while stage 2 leverages input attention masking to restrict its receptive field to these regions, enabling a focused analysis while filtering out potentially spurious information. Both stages are trained jointly, allowing stage 2 to refine stage 1. Extensive experiments across diverse benchmarks demonstrate that our approach significantly improves robustness against spurious correlations and out-of-distribution backgrounds.

This paper studies a conceptually simple extension of Masked Autoencoders (MAE) to spatiotemporal representation learning from videos. We randomly mask out spacetime patches in videos and learn an autoencoder to reconstruct them in pixels. Interestingly, we show that our MAE method can learn strong representations with almost no inductive bias on spacetime (only except for patch and positional embeddings), and spacetime-agnostic random masking performs the best. We observe that the optimal masking ratio is as high as 90% (vs. 75% on images), supporting the hypothesis that this ratio is related to information redundancy of the data. A high masking ratio leads to a large speedup, e.g., > 4x in wall-clock time or even more. We report competitive results on several challenging video datasets using vanilla Vision Transformers. We observe that MAE can outperform supervised pre-training by large margins. We further report encouraging results of training on real-world, uncurated Instagram data. Our study suggests that the general framework of masked autoencoding (BERT, MAE, etc.) can be a unified methodology for representation learning with minimal domain knowledge.

Predicting and anticipating future outcomes or reasoning about missing information in a sequence are critical skills for agents to be able to make intelligent decisions. This requires strong, temporally coherent generative capabilities. Diffusion models have shown remarkable success in several generative tasks, but have not been extensively explored in the video domain. We present Random-Mask Video Diffusion (RaMViD), which extends image diffusion models to videos using 3D convolutions, and introduces a new conditioning technique during training. By varying the mask we condition on, the model is able to perform video prediction, infilling, and upsampling. Due to our simple conditioning scheme, we can utilize the same architecture as used for unconditional training, which allows us to train the model in a conditional and unconditional fashion at the same time. We evaluate RaMViD on two benchmark datasets for video prediction, on which we achieve state-of-the-art results, and one for video generation. High-resolution videos are provided at https://sites.google.com/view/video-diffusion-prediction.

We present an approach to pose object recognition as next token prediction. The idea is to apply a language decoder that auto-regressively predicts the text tokens from image embeddings to form labels. To ground this prediction process in auto-regression, we customize a non-causal attention mask for the decoder, incorporating two key features: modeling tokens from different labels to be independent, and treating image tokens as a prefix. This masking mechanism inspires an efficient method - one-shot sampling - to simultaneously sample tokens of multiple labels in parallel and rank generated labels by their probabilities during inference. To further enhance the efficiency, we propose a simple strategy to construct a compact decoder by simply discarding the intermediate blocks of a pretrained language model. This approach yields a decoder that matches the full model's performance while being notably more efficient. The code is available at https://github.com/kaiyuyue/nxtp

Masked Image Modeling (MIM) has recently been established as a potent pre-training paradigm. A pretext task is constructed by masking patches in an input image, and this masked content is then predicted by a neural network using visible patches as sole input. This pre-training leads to state-of-the-art performance when finetuned for high-level semantic tasks, e.g. image classification and object detection. In this paper we instead seek to learn representations that transfer well to a wide variety of 3D vision and lower-level geometric downstream tasks, such as depth prediction or optical flow estimation. Inspired by MIM, we propose an unsupervised representation learning task trained from pairs of images showing the same scene from different viewpoints. More precisely, we propose the pretext task of cross-view completion where the first input image is partially masked, and this masked content has to be reconstructed from the visible content and the second image. In single-view MIM, the masked content often cannot be inferred precisely from the visible portion only, so the model learns to act as a prior influenced by high-level semantics. In contrast, this ambiguity can be resolved with cross-view completion from the second unmasked image, on the condition that the model is able to understand the spatial relationship between the two images. Our experiments show that our pretext task leads to significantly improved performance for monocular 3D vision downstream tasks such as depth estimation. In addition, our model can be directly applied to binocular downstream tasks like optical flow or relative camera pose estimation, for which we obtain competitive results without bells and whistles, i.e., using a generic architecture without any task-specific design.

A great challenge in video-language (VidL) modeling lies in the disconnection between fixed video representations extracted from image/video understanding models and downstream VidL data. Recent studies try to mitigate this disconnection via end-to-end training. To make it computationally feasible, prior works tend to "imagify" video inputs, i.e., a handful of sparsely sampled frames are fed into a 2D CNN, followed by a simple mean-pooling or concatenation to obtain the overall video representations. Although achieving promising results, such simple approaches may lose temporal information that is essential for performing downstream VidL tasks. In this work, we present VIOLET, a fully end-to-end VIdeO-LanguagE Transformer, which adopts a video transformer to explicitly model the temporal dynamics of video inputs. Further, unlike previous studies that found pre-training tasks on video inputs (e.g., masked frame modeling) not very effective, we design a new pre-training task, Masked Visual-token Modeling (MVM), for better video modeling. Specifically, the original video frame patches are "tokenized" into discrete visual tokens, and the goal is to recover the original visual tokens based on the masked patches. Comprehensive analysis demonstrates the effectiveness of both explicit temporal modeling via video transformer and MVM. As a result, VIOLET achieves new state-of-the-art performance on 5 video question answering tasks and 4 text-to-video retrieval tasks.

Existing video generation models struggle to follow complex text prompts and synthesize multiple objects, raising the need for additional grounding input for improved controllability. In this work, we propose to decompose videos into visual primitives - blob video representation, a general representation for controllable video generation. Based on blob conditions, we develop a blob-grounded video diffusion model named BlobGEN-Vid that allows users to control object motions and fine-grained object appearance. In particular, we introduce a masked 3D attention module that effectively improves regional consistency across frames. In addition, we introduce a learnable module to interpolate text embeddings so that users can control semantics in specific frames and obtain smooth object transitions. We show that our framework is model-agnostic and build BlobGEN-Vid based on both U-Net and DiT-based video diffusion models. Extensive experimental results show that BlobGEN-Vid achieves superior zero-shot video generation ability and state-of-the-art layout controllability on multiple benchmarks. When combined with an LLM for layout planning, our framework even outperforms proprietary text-to-video generators in terms of compositional accuracy.

We present a conceptually simple, flexible, and general framework for object instance segmentation. Our approach efficiently detects objects in an image while simultaneously generating a high-quality segmentation mask for each instance. The method, called Mask R-CNN, extends Faster R-CNN by adding a branch for predicting an object mask in parallel with the existing branch for bounding box recognition. Mask R-CNN is simple to train and adds only a small overhead to Faster R-CNN, running at 5 fps. Moreover, Mask R-CNN is easy to generalize to other tasks, e.g., allowing us to estimate human poses in the same framework. We show top results in all three tracks of the COCO suite of challenges, including instance segmentation, bounding-box object detection, and person keypoint detection. Without bells and whistles, Mask R-CNN outperforms all existing, single-model entries on every task, including the COCO 2016 challenge winners. We hope our simple and effective approach will serve as a solid baseline and help ease future research in instance-level recognition. Code has been made available at: https://github.com/facebookresearch/Detectron

We present a novel masked image modeling (MIM) approach, context autoencoder (CAE), for self-supervised representation pretraining. We pretrain an encoder by making predictions in the encoded representation space. The pretraining tasks include two tasks: masked representation prediction - predict the representations for the masked patches, and masked patch reconstruction - reconstruct the masked patches. The network is an encoder-regressor-decoder architecture: the encoder takes the visible patches as input; the regressor predicts the representations of the masked patches, which are expected to be aligned with the representations computed from the encoder, using the representations of visible patches and the positions of visible and masked patches; the decoder reconstructs the masked patches from the predicted encoded representations. The CAE design encourages the separation of learning the encoder (representation) from completing the pertaining tasks: masked representation prediction and masked patch reconstruction tasks, and making predictions in the encoded representation space empirically shows the benefit to representation learning. We demonstrate the effectiveness of our CAE through superior transfer performance in downstream tasks: semantic segmentation, object detection and instance segmentation, and classification. The code will be available at https://github.com/Atten4Vis/CAE.

Masked Image Modeling (MIM) has emerged as a promising method for deriving visual representations from unlabeled image data by predicting missing pixels from masked portions of images. It excels in region-aware learning and provides strong initializations for various tasks, but struggles to capture high-level semantics without further supervised fine-tuning, likely due to the low-level nature of its pixel reconstruction objective. A promising yet unrealized framework is learning representations through masked reconstruction in latent space, combining the locality of MIM with the high-level targets. However, this approach poses significant training challenges as the reconstruction targets are learned in conjunction with the model, potentially leading to trivial or suboptimal solutions.Our study is among the first to thoroughly analyze and address the challenges of such framework, which we refer to as Latent MIM. Through a series of carefully designed experiments and extensive analysis, we identify the source of these challenges, including representation collapsing for joint online/target optimization, learning objectives, the high region correlation in latent space and decoding conditioning. By sequentially addressing these issues, we demonstrate that Latent MIM can indeed learn high-level representations while retaining the benefits of MIM models.

Despite the prevailing transition from single-task to multi-task approaches in video anomaly detection, we observe that many adopt sub-optimal frameworks for individual proxy tasks. Motivated by this, we contend that optimizing single-task frameworks can advance both single- and multi-task approaches. Accordingly, we leverage middle-frame prediction as the primary proxy task, and introduce an effective hybrid framework designed to generate accurate predictions for normal frames and flawed predictions for abnormal frames. This hybrid framework is built upon a bi-directional structure that seamlessly integrates both vision transformers and ConvLSTMs. Specifically, we utilize this bi-directional structure to fully analyze the temporal dimension by predicting frames in both forward and backward directions, significantly boosting the detection stability. Given the transformer's capacity to model long-range contextual dependencies, we develop a convolutional temporal transformer that efficiently associates feature maps from all context frames to generate attention-based predictions for target frames. Furthermore, we devise a layer-interactive ConvLSTM bridge that facilitates the smooth flow of low-level features across layers and time-steps, thereby strengthening predictions with fine details. Anomalies are eventually identified by scrutinizing the discrepancies between target frames and their corresponding predictions. Several experiments conducted on public benchmarks affirm the efficacy of our hybrid framework, whether used as a standalone single-task approach or integrated as a branch in a multi-task approach. These experiments also underscore the advantages of merging vision transformers and ConvLSTMs for video anomaly detection.

We present an unsupervised visual feature learning algorithm driven by context-based pixel prediction. By analogy with auto-encoders, we propose Context Encoders -- a convolutional neural network trained to generate the contents of an arbitrary image region conditioned on its surroundings. In order to succeed at this task, context encoders need to both understand the content of the entire image, as well as produce a plausible hypothesis for the missing part(s). When training context encoders, we have experimented with both a standard pixel-wise reconstruction loss, as well as a reconstruction plus an adversarial loss. The latter produces much sharper results because it can better handle multiple modes in the output. We found that a context encoder learns a representation that captures not just appearance but also the semantics of visual structures. We quantitatively demonstrate the effectiveness of our learned features for CNN pre-training on classification, detection, and segmentation tasks. Furthermore, context encoders can be used for semantic inpainting tasks, either stand-alone or as initialization for non-parametric methods.

In this paper, we explore the visual representations produced from a pre-trained text-to-video (T2V) diffusion model for video understanding tasks. We hypothesize that the latent representation learned from a pretrained generative T2V model encapsulates rich semantics and coherent temporal correspondences, thereby naturally facilitating video understanding. Our hypothesis is validated through the classic referring video object segmentation (R-VOS) task. We introduce a novel framework, termed "VD-IT", tailored with dedicatedly designed components built upon a fixed pretrained T2V model. Specifically, VD-IT uses textual information as a conditional input, ensuring semantic consistency across time for precise temporal instance matching. It further incorporates image tokens as supplementary textual inputs, enriching the feature set to generate detailed and nuanced masks. Besides, instead of using the standard Gaussian noise, we propose to predict the video-specific noise with an extra noise prediction module, which can help preserve the feature fidelity and elevates segmentation quality. Through extensive experiments, we surprisingly observe that fixed generative T2V diffusion models, unlike commonly used video backbones (e.g., Video Swin Transformer) pretrained with discriminative image/video pre-tasks, exhibit better potential to maintain semantic alignment and temporal consistency. On existing standard benchmarks, our VD-IT achieves highly competitive results, surpassing many existing state-of-the-art methods. The code is available at https://github.com/buxiangzhiren/VD-IT.

We present HERO, a novel framework for large-scale video+language omni-representation learning. HERO encodes multimodal inputs in a hierarchical structure, where local context of a video frame is captured by a Cross-modal Transformer via multimodal fusion, and global video context is captured by a Temporal Transformer. In addition to standard Masked Language Modeling (MLM) and Masked Frame Modeling (MFM) objectives, we design two new pre-training tasks: (i) Video-Subtitle Matching (VSM), where the model predicts both global and local temporal alignment; and (ii) Frame Order Modeling (FOM), where the model predicts the right order of shuffled video frames. HERO is jointly trained on HowTo100M and large-scale TV datasets to gain deep understanding of complex social dynamics with multi-character interactions. Comprehensive experiments demonstrate that HERO achieves new state of the art on multiple benchmarks over Text-based Video/Video-moment Retrieval, Video Question Answering (QA), Video-and-language Inference and Video Captioning tasks across different domains. We also introduce two new challenging benchmarks How2QA and How2R for Video QA and Retrieval, collected from diverse video content over multimodalities.

Masked autoencoders have become popular training paradigms for self-supervised visual representation learning. These models randomly mask a portion of the input and reconstruct the masked portion according to the target representations. In this paper, we first show that a careful choice of the target representation is unnecessary for learning good representations, since different targets tend to derive similarly behaved models. Driven by this observation, we propose a multi-stage masked distillation pipeline and use a randomly initialized model as the teacher, enabling us to effectively train high-capacity models without any efforts to carefully design target representations. Interestingly, we further explore using teachers of larger capacity, obtaining distilled students with remarkable transferring ability. On different tasks of classification, transfer learning, object detection, and semantic segmentation, the proposed method to perform masked knowledge distillation with bootstrapped teachers (dBOT) outperforms previous self-supervised methods by nontrivial margins. We hope our findings, as well as the proposed method, could motivate people to rethink the roles of target representations in pre-training masked autoencoders.The code and pre-trained models are publicly available at https://github.com/liuxingbin/dbot.

Current video understanding models rely on fixed frame sampling strategies, processing predetermined visual inputs regardless of the specific reasoning requirements of each question. This static approach limits their ability to adaptively gather visual evidence, leading to suboptimal performance on tasks that require either broad temporal coverage or fine-grained spatial detail. In this paper, we introduce FrameMind, an end-to-end framework trained with reinforcement learning that enables models to dynamically request visual information during reasoning through Frame-Interleaved Chain-of-Thought (FiCOT). Unlike traditional approaches, FrameMind operates in multiple turns where the model alternates between textual reasoning and active visual perception, using tools to extract targeted frames or video clips based on identified knowledge gaps. To train effective dynamic sampling policies, we propose Dynamic Resolution Frame Sampling (DRFS), which exposes models to diverse temporal-spatial trade-offs during learning, and DRFS-GRPO, a group-relative policy optimization algorithm that learns from outcome-based rewards without requiring frame-level annotations. Extensive experiments on challenging benchmarks like MLVU and VideoMME demonstrate that our method significantly outperforms existing models, advancing the state of the art in flexible and efficient video understanding.

Video DiTs have advanced video generation, yet they still struggle to model multi-instance or subject-object interactions. This raises a key question: How do these models internally represent interactions? To answer this, we curate MATRIX-11K, a video dataset with interaction-aware captions and multi-instance mask tracks. Using this dataset, we conduct a systematic analysis that formalizes two perspectives of video DiTs: semantic grounding, via video-to-text attention, which evaluates whether noun and verb tokens capture instances and their relations; and semantic propagation, via video-to-video attention, which assesses whether instance bindings persist across frames. We find both effects concentrate in a small subset of interaction-dominant layers. Motivated by this, we introduce MATRIX, a simple and effective regularization that aligns attention in specific layers of video DiTs with multi-instance mask tracks from the MATRIX-11K dataset, enhancing both grounding and propagation. We further propose InterGenEval, an evaluation protocol for interaction-aware video generation. In experiments, MATRIX improves both interaction fidelity and semantic alignment while reducing drift and hallucination. Extensive ablations validate our design choices. Codes and weights will be released.

We present REM, a framework for segmenting a wide range of concepts in video that can be described through natural language. Our method capitalizes on visual-language representations learned by video diffusion models on Internet-scale datasets. A key insight of our approach is preserving as much of the generative model's original representation as possible, while fine-tuning it on narrow-domain Referral Object Segmentation datasets. As a result, our framework can accurately segment and track rare and unseen objects, despite being trained on object masks from a limited set of categories. Additionally, it can generalize to non-object dynamic concepts, such as waves crashing in the ocean, as demonstrated in our newly introduced benchmark for Referral Video Process Segmentation (Ref-VPS). Our experiments show that REM performs on par with state-of-the-art approaches on in-domain datasets, like Ref-DAVIS, while outperforming them by up to twelve points in terms of region similarity on out-of-domain data, leveraging the power of Internet-scale pre-training.

Token-based masked generative models are gaining popularity for their fast inference time with parallel decoding. While recent token-based approaches achieve competitive performance to diffusion-based models, their generation performance is still suboptimal as they sample multiple tokens simultaneously without considering the dependence among them. We empirically investigate this problem and propose a learnable sampling model, Text-Conditioned Token Selection (TCTS), to select optimal tokens via localized supervision with text information. TCTS improves not only the image quality but also the semantic alignment of the generated images with the given texts. To further improve the image quality, we introduce a cohesive sampling strategy, Frequency Adaptive Sampling (FAS), to each group of tokens divided according to the self-attention maps. We validate the efficacy of TCTS combined with FAS with various generative tasks, demonstrating that it significantly outperforms the baselines in image-text alignment and image quality. Our text-conditioned sampling framework further reduces the original inference time by more than 50% without modifying the original generative model.

When automatically generating a sentence description for an image or video, it often remains unclear how well the generated caption is grounded, that is whether the model uses the correct image regions to output particular words, or if the model is hallucinating based on priors in the dataset and/or the language model. The most common way of relating image regions with words in caption models is through an attention mechanism over the regions that are used as input to predict the next word. The model must therefore learn to predict the attentional weights without knowing the word it should localize. This is difficult to train without grounding supervision since recurrent models can propagate past information and there is no explicit signal to force the captioning model to properly ground the individual decoded words. In this work, we help the model to achieve this via a novel cyclical training regimen that forces the model to localize each word in the image after the sentence decoder generates it, and then reconstruct the sentence from the localized image region(s) to match the ground-truth. Our proposed framework only requires learning one extra fully-connected layer (the localizer), a layer that can be removed at test time. We show that our model significantly improves grounding accuracy without relying on grounding supervision or introducing extra computation during inference, for both image and video captioning tasks. Code is available at https://github.com/chihyaoma/cyclical-visual-captioning .

Vision Transformers (ViTs) have delivered remarkable progress through global self-attention, yet their quadratic complexity can become prohibitive for high-resolution inputs. In this work, we present ViT-Linearizer, a cross-architecture distillation framework that transfers rich ViT representations into a linear-time, recurrent-style model. Our approach leverages 1) activation matching, an intermediate constraint that encourages student to align its token-wise dependencies with those produced by the teacher, and 2) masked prediction, a contextual reconstruction objective that requires the student to predict the teacher's representations for unseen (masked) tokens, to effectively distill the quadratic self-attention knowledge into the student while maintaining efficient complexity. Empirically, our method provides notable speedups particularly for high-resolution tasks, significantly addressing the hardware challenges in inference. Additionally, it also elevates Mamba-based architectures' performance on standard vision benchmarks, achieving a competitive 84.3% top-1 accuracy on ImageNet with a base-sized model. Our results underscore the good potential of RNN-based solutions for large-scale visual tasks, bridging the gap between theoretical efficiency and real-world practice.

In this paper, we propose a novel video depth estimation approach, FutureDepth, which enables the model to implicitly leverage multi-frame and motion cues to improve depth estimation by making it learn to predict the future at training. More specifically, we propose a future prediction network, F-Net, which takes the features of multiple consecutive frames and is trained to predict multi-frame features one time step ahead iteratively. In this way, F-Net learns the underlying motion and correspondence information, and we incorporate its features into the depth decoding process. Additionally, to enrich the learning of multiframe correspondence cues, we further leverage a reconstruction network, R-Net, which is trained via adaptively masked auto-encoding of multiframe feature volumes. At inference time, both F-Net and R-Net are used to produce queries to work with the depth decoder, as well as a final refinement network. Through extensive experiments on several benchmarks, i.e., NYUDv2, KITTI, DDAD, and Sintel, which cover indoor, driving, and open-domain scenarios, we show that FutureDepth significantly improves upon baseline models, outperforms existing video depth estimation methods, and sets new state-of-the-art (SOTA) accuracy. Furthermore, FutureDepth is more efficient than existing SOTA video depth estimation models and has similar latencies when comparing to monocular models

We propose Masked Siamese Networks (MSN), a self-supervised learning framework for learning image representations. Our approach matches the representation of an image view containing randomly masked patches to the representation of the original unmasked image. This self-supervised pre-training strategy is particularly scalable when applied to Vision Transformers since only the unmasked patches are processed by the network. As a result, MSNs improve the scalability of joint-embedding architectures, while producing representations of a high semantic level that perform competitively on low-shot image classification. For instance, on ImageNet-1K, with only 5,000 annotated images, our base MSN model achieves 72.4% top-1 accuracy, and with 1% of ImageNet-1K labels, we achieve 75.7% top-1 accuracy, setting a new state-of-the-art for self-supervised learning on this benchmark. Our code is publicly available.

Referring video object segmentation (RVOS) is a task that aims to segment the target object in all video frames based on a sentence describing the object. Previous RVOS methods have achieved significant performance with densely-annotated datasets, whose construction is expensive and time-consuming. To relieve the burden of data annotation while maintaining sufficient supervision for segmentation, we propose a new annotation scheme, in which we label the frame where the object first appears with a mask and use bounding boxes for the subsequent frames. Based on this scheme, we propose a method to learn from this weak annotation. Specifically, we design a cross frame segmentation method, which uses the language-guided dynamic filters to thoroughly leverage the valuable mask annotation and bounding boxes. We further develop a bi-level contrastive learning method to encourage the model to learn discriminative representation at the pixel level. Extensive experiments and ablative analyses show that our method is able to achieve competitive performance without the demand of dense mask annotation. The code will be available at https://github.com/wangbo-zhao/WRVOS/.

Unsupervised domain adaptation (UDA) for semantic segmentation aims to transfer models from a labeled source domain to an unlabeled target domain. While auxiliary self-supervised tasks-particularly contrastive learning-have improved feature discriminability, masked modeling approaches remain underexplored in this setting, largely due to architectural incompatibility and misaligned optimization objectives. We propose Masked Feature Modeling (MFM), a novel auxiliary task that performs feature masking and reconstruction directly in the feature space. Unlike existing masked modeling methods that reconstruct low-level inputs or perceptual features (e.g., HOG or visual tokens), MFM aligns its learning target with the main segmentation task, ensuring compatibility with standard architectures like DeepLab and DAFormer without modifying the inference pipeline. To facilitate effective reconstruction, we introduce a lightweight auxiliary module, Rebuilder, which is trained jointly but discarded during inference, adding zero computational overhead at test time. Crucially, MFM leverages the segmentation decoder to classify the reconstructed features, tightly coupling the auxiliary objective with the pixel-wise prediction task to avoid interference with the primary task. Extensive experiments across various architectures and UDA benchmarks demonstrate that MFM consistently enhances segmentation performance, offering a simple, efficient, and generalizable strategy for unsupervised domain-adaptive semantic segmentation.

Masked visual modeling has attracted much attention due to its promising potential in learning generalizable representations. Typical approaches urge models to predict specific contents of masked tokens, which can be intuitively considered as teaching a student (the model) to solve given problems (predicting masked contents). Under such settings, the performance is highly correlated with mask strategies (the difficulty of provided problems). We argue that it is equally important for the model to stand in the shoes of a teacher to produce challenging problems by itself. Intuitively, patches with high values of reconstruction loss can be regarded as hard samples, and masking those hard patches naturally becomes a demanding reconstruction task. To empower the model as a teacher, we propose Hard Patches Mining (HPM), predicting patch-wise losses and subsequently determining where to mask. Technically, we introduce an auxiliary loss predictor, which is trained with a relative objective to prevent overfitting to exact loss values. Also, to gradually guide the training procedure, we propose an easy-to-hard mask strategy. Empirically, HPM brings significant improvements under both image and video benchmarks. Interestingly, solely incorporating the extra loss prediction objective leads to better representations, verifying the efficacy of determining where is hard to reconstruct. The code is available at https://github.com/Haochen-Wang409/HPM.

The remarkable efficacy of text-to-image diffusion models has motivated extensive exploration of their potential application in video domains. Zero-shot methods seek to extend image diffusion models to videos without necessitating model training. Recent methods mainly focus on incorporating inter-frame correspondence into attention mechanisms. However, the soft constraint imposed on determining where to attend to valid features can sometimes be insufficient, resulting in temporal inconsistency. In this paper, we introduce FRESCO, intra-frame correspondence alongside inter-frame correspondence to establish a more robust spatial-temporal constraint. This enhancement ensures a more consistent transformation of semantically similar content across frames. Beyond mere attention guidance, our approach involves an explicit update of features to achieve high spatial-temporal consistency with the input video, significantly improving the visual coherence of the resulting translated videos. Extensive experiments demonstrate the effectiveness of our proposed framework in producing high-quality, coherent videos, marking a notable improvement over existing zero-shot methods.

Prompt tuning and adapter tuning have shown great potential in transferring pre-trained vision-language models (VLMs) to various downstream tasks. In this work, we design a new type of tuning method, termed as regularized mask tuning, which masks the network parameters through a learnable selection. Inspired by neural pathways, we argue that the knowledge required by a downstream task already exists in the pre-trained weights but just gets concealed in the upstream pre-training stage. To bring the useful knowledge back into light, we first identify a set of parameters that are important to a given downstream task, then attach a binary mask to each parameter, and finally optimize these masks on the downstream data with the parameters frozen. When updating the mask, we introduce a novel gradient dropout strategy to regularize the parameter selection, in order to prevent the model from forgetting old knowledge and overfitting the downstream data. Experimental results on 11 datasets demonstrate the consistent superiority of our method over previous alternatives. It is noteworthy that we manage to deliver 18.73% performance improvement compared to the zero-shot CLIP via masking an average of only 2.56% parameters. Furthermore, our method is synergistic with most existing parameter-efficient tuning methods and can boost the performance on top of them. Project page can be found here (https://wuw2019.github.io/R-AMT/).

End-to-end training from scratch of current deep architectures for new computer vision problems would require Imagenet-scale datasets, and this is not always possible. In this paper we present a method that is able to take advantage of freely available multi-modal content to train computer vision algorithms without human supervision. We put forward the idea of performing self-supervised learning of visual features by mining a large scale corpus of multi-modal (text and image) documents. We show that discriminative visual features can be learnt efficiently by training a CNN to predict the semantic context in which a particular image is more probable to appear as an illustration. For this we leverage the hidden semantic structures discovered in the text corpus with a well-known topic modeling technique. Our experiments demonstrate state of the art performance in image classification, object detection, and multi-modal retrieval compared to recent self-supervised or natural-supervised approaches.

Masked (or absorbing) diffusion is actively explored as an alternative to autoregressive models for generative modeling of discrete data. However, existing work in this area has been hindered by unnecessarily complex model formulations and unclear relationships between different perspectives, leading to suboptimal parameterization, training objectives, and ad hoc adjustments to counteract these issues. In this work, we aim to provide a simple and general framework that unlocks the full potential of masked diffusion models. We show that the continuous-time variational objective of masked diffusion models is a simple weighted integral of cross-entropy losses. Our framework also enables training generalized masked diffusion models with state-dependent masking schedules. When evaluated by perplexity, our models trained on OpenWebText surpass prior diffusion language models at GPT-2 scale and demonstrate superior performance on 4 out of 5 zero-shot language modeling tasks. Furthermore, our models vastly outperform previous discrete diffusion models on pixel-level image modeling, achieving 2.78~(CIFAR-10) and 3.42 (ImageNet 64times64) bits per dimension that are comparable or better than autoregressive models of similar sizes.

In this paper, we introduce Positional Skip-wisE (PoSE) training for efficient adaptation of large language models~(LLMs) to extremely long context windows. PoSE decouples train length from target context window size by simulating long inputs using a fixed context window with manipulated position indices during training. Concretely, we select several short chunks from a long input sequence, and introduce distinct skipping bias terms to modify the position indices of each chunk. These bias terms, along with the length of each chunk, are altered for each training example, allowing the model to adapt to all positions within the target context window without training on full length inputs. Experiments show that, compared with fine-tuning on the full length, PoSE greatly reduces memory and time overhead with minimal impact on performance. Leveraging this advantage, we have successfully extended the LLaMA model to 128k tokens. Furthermore, we empirically confirm that PoSE is compatible with all RoPE-based LLMs and various position interpolation strategies. Notably, by decoupling fine-tuning length from target context window, PoSE can theoretically extend the context window infinitely, constrained only by memory usage for inference. With ongoing advancements for efficient inference, we believe PoSE holds great promise for scaling the context window even further.

Inspired by recent work in machine translation and object detection, we introduce an attention based model that automatically learns to describe the content of images. We describe how we can train this model in a deterministic manner using standard backpropagation techniques and stochastically by maximizing a variational lower bound. We also show through visualization how the model is able to automatically learn to fix its gaze on salient objects while generating the corresponding words in the output sequence. We validate the use of attention with state-of-the-art performance on three benchmark datasets: Flickr8k, Flickr30k and MS COCO.

Recent work has shown great progress in integrating spatial conditioning to control large, pre-trained text-to-image diffusion models. Despite these advances, existing methods describe the spatial image content using hand-crafted conditioning inputs, which are either semantically ambiguous (e.g., edges) or require expensive manual annotations (e.g., semantic segmentation). To address these limitations, we propose a new label-free way of conditioning diffusion models to enable fine-grained spatial control. We introduce the concept of neural semantic image synthesis, which uses neural layouts extracted from pre-trained foundation models as conditioning. Neural layouts are advantageous as they provide rich descriptions of the desired image, containing both semantics and detailed geometry of the scene. We experimentally show that images synthesized via neural semantic image synthesis achieve similar or superior pixel-level alignment of semantic classes compared to those created using expensive semantic label maps. At the same time, they capture better semantics, instance separation, and object orientation than other label-free conditioning options, such as edges or depth. Moreover, we show that images generated by neural layout conditioning can effectively augment real data for training various perception tasks.

Masked Autoencoders (MAEs) learn generalizable representations for image, text, audio, video, etc., by reconstructing masked input data from tokens of the visible data. Current MAE approaches for videos rely on random patch, tube, or frame-based masking strategies to select these tokens. This paper proposes AdaMAE, an adaptive masking strategy for MAEs that is end-to-end trainable. Our adaptive masking strategy samples visible tokens based on the semantic context using an auxiliary sampling network. This network estimates a categorical distribution over spacetime-patch tokens. The tokens that increase the expected reconstruction error are rewarded and selected as visible tokens, motivated by the policy gradient algorithm in reinforcement learning. We show that AdaMAE samples more tokens from the high spatiotemporal information regions, thereby allowing us to mask 95% of tokens, resulting in lower memory requirements and faster pre-training. We conduct ablation studies on the Something-Something v2 (SSv2) dataset to demonstrate the efficacy of our adaptive sampling approach and report state-of-the-art results of 70.0% and 81.7% in top-1 accuracy on SSv2 and Kinetics-400 action classification datasets with a ViT-Base backbone and 800 pre-training epochs.

Effective modeling of complex spatiotemporal dependencies in long-form videos remains an open problem. The recently proposed Structured State-Space Sequence (S4) model with its linear complexity offers a promising direction in this space. However, we demonstrate that treating all image-tokens equally as done by S4 model can adversely affect its efficiency and accuracy. To address this limitation, we present a novel Selective S4 (i.e., S5) model that employs a lightweight mask generator to adaptively select informative image tokens resulting in more efficient and accurate modeling of long-term spatiotemporal dependencies in videos. Unlike previous mask-based token reduction methods used in transformers, our S5 model avoids the dense self-attention calculation by making use of the guidance of the momentum-updated S4 model. This enables our model to efficiently discard less informative tokens and adapt to various long-form video understanding tasks more effectively. However, as is the case for most token reduction methods, the informative image tokens could be dropped incorrectly. To improve the robustness and the temporal horizon of our model, we propose a novel long-short masked contrastive learning (LSMCL) approach that enables our model to predict longer temporal context using shorter input videos. We present extensive comparative results using three challenging long-form video understanding datasets (LVU, COIN and Breakfast), demonstrating that our approach consistently outperforms the previous state-of-the-art S4 model by up to 9.6% accuracy while reducing its memory footprint by 23%.

High-quality video inpainting that completes missing regions in video frames is a promising yet challenging task. State-of-the-art approaches adopt attention models to complete a frame by searching missing contents from reference frames, and further complete whole videos frame by frame. However, these approaches can suffer from inconsistent attention results along spatial and temporal dimensions, which often leads to blurriness and temporal artifacts in videos. In this paper, we propose to learn a joint Spatial-Temporal Transformer Network (STTN) for video inpainting. Specifically, we simultaneously fill missing regions in all input frames by self-attention, and propose to optimize STTN by a spatial-temporal adversarial loss. To show the superiority of the proposed model, we conduct both quantitative and qualitative evaluations by using standard stationary masks and more realistic moving object masks. Demo videos are available at https://github.com/researchmm/STTN.

This paper presents a simple yet effective approach to modeling space-time correspondences in the context of video object segmentation. Unlike most existing approaches, we establish correspondences directly between frames without re-encoding the mask features for every object, leading to a highly efficient and robust framework. With the correspondences, every node in the current query frame is inferred by aggregating features from the past in an associative fashion. We cast the aggregation process as a voting problem and find that the existing inner-product affinity leads to poor use of memory with a small (fixed) subset of memory nodes dominating the votes, regardless of the query. In light of this phenomenon, we propose using the negative squared Euclidean distance instead to compute the affinities. We validated that every memory node now has a chance to contribute, and experimentally showed that such diversified voting is beneficial to both memory efficiency and inference accuracy. The synergy of correspondence networks and diversified voting works exceedingly well, achieves new state-of-the-art results on both DAVIS and YouTubeVOS datasets while running significantly faster at 20+ FPS for multiple objects without bells and whistles.

Understanding causal event relationships and achieving fine-grained temporal grounding in videos remain challenging for vision-language models. Existing methods either compress video tokens to reduce temporal resolution, or treat videos as unsegmented streams, which obscures fine-grained event boundaries and limits the modeling of causal dependencies. We propose TEMPURA (Temporal Event Masked Prediction and Understanding for Reasoning in Action), a two-stage training framework that enhances video temporal understanding. TEMPURA first applies masked event prediction reasoning to reconstruct missing events and generate step-by-step causal explanations from dense event annotations, drawing inspiration from effective infilling techniques. TEMPURA then learns to perform video segmentation and dense captioning to decompose videos into non-overlapping events with detailed, timestamp-aligned descriptions. We train TEMPURA on VER, a large-scale dataset curated by us that comprises 1M training instances and 500K videos with temporally aligned event descriptions and structured reasoning steps. Experiments on temporal grounding and highlight detection benchmarks demonstrate that TEMPURA outperforms strong baseline models, confirming that integrating causal reasoning with fine-grained temporal segmentation leads to improved video understanding.

We aim to investigate whether end-to-end learning of visual reasoning can be achieved with general-purpose neural networks, with the help of visual pretraining. A positive result would refute the common belief that explicit visual abstraction (e.g. object detection) is essential for compositional generalization on visual reasoning, and confirm the feasibility of a neural network "generalist" to solve visual recognition and reasoning tasks. We propose a simple and general self-supervised framework which "compresses" each video frame into a small set of tokens with a transformer network, and reconstructs the remaining frames based on the compressed temporal context. To minimize the reconstruction loss, the network must learn a compact representation for each image, as well as capture temporal dynamics and object permanence from temporal context. We perform evaluation on two visual reasoning benchmarks, CATER and ACRE. We observe that pretraining is essential to achieve compositional generalization for end-to-end visual reasoning. Our proposed framework outperforms traditional supervised pretraining, including image classification and explicit object detection, by large margins.

Recent advances in customized video generation have enabled users to create videos tailored to both specific subjects and motion trajectories. However, existing methods often require complicated test-time fine-tuning and struggle with balancing subject learning and motion control, limiting their real-world applications. In this paper, we present DreamVideo-2, a zero-shot video customization framework capable of generating videos with a specific subject and motion trajectory, guided by a single image and a bounding box sequence, respectively, and without the need for test-time fine-tuning. Specifically, we introduce reference attention, which leverages the model's inherent capabilities for subject learning, and devise a mask-guided motion module to achieve precise motion control by fully utilizing the robust motion signal of box masks derived from bounding boxes. While these two components achieve their intended functions, we empirically observe that motion control tends to dominate over subject learning. To address this, we propose two key designs: 1) the masked reference attention, which integrates a blended latent mask modeling scheme into reference attention to enhance subject representations at the desired positions, and 2) a reweighted diffusion loss, which differentiates the contributions of regions inside and outside the bounding boxes to ensure a balance between subject and motion control. Extensive experimental results on a newly curated dataset demonstrate that DreamVideo-2 outperforms state-of-the-art methods in both subject customization and motion control. The dataset, code, and models will be made publicly available.

We consider the task of Image-to-Video (I2V) generation, which involves transforming static images into realistic video sequences based on a textual description. While recent advancements produce photorealistic outputs, they frequently struggle to create videos with accurate and consistent object motion, especially in multi-object scenarios. To address these limitations, we propose a two-stage compositional framework that decomposes I2V generation into: (i) An explicit intermediate representation generation stage, followed by (ii) A video generation stage that is conditioned on this representation. Our key innovation is the introduction of a mask-based motion trajectory as an intermediate representation, that captures both semantic object information and motion, enabling an expressive but compact representation of motion and semantics. To incorporate the learned representation in the second stage, we utilize object-level attention objectives. Specifically, we consider a spatial, per-object, masked-cross attention objective, integrating object-specific prompts into corresponding latent space regions and a masked spatio-temporal self-attention objective, ensuring frame-to-frame consistency for each object. We evaluate our method on challenging benchmarks with multi-object and high-motion scenarios and empirically demonstrate that the proposed method achieves state-of-the-art results in temporal coherence, motion realism, and text-prompt faithfulness. Additionally, we introduce \benchmark, a new challenging benchmark for single-object and multi-object I2V generation, and demonstrate our method's superiority on this benchmark. Project page is available at https://guyyariv.github.io/TTM/.

We introduce MAGI, a hybrid video generation framework that combines masked modeling for intra-frame generation with causal modeling for next-frame generation. Our key innovation, Complete Teacher Forcing (CTF), conditions masked frames on complete observation frames rather than masked ones (namely Masked Teacher Forcing, MTF), enabling a smooth transition from token-level (patch-level) to frame-level autoregressive generation. CTF significantly outperforms MTF, achieving a +23% improvement in FVD scores on first-frame conditioned video prediction. To address issues like exposure bias, we employ targeted training strategies, setting a new benchmark in autoregressive video generation. Experiments show that MAGI can generate long, coherent video sequences exceeding 100 frames, even when trained on as few as 16 frames, highlighting its potential for scalable, high-quality video generation.

Masked Image Modeling (MIM) arises as a promising option for Vision Transformers among various self-supervised learning (SSL) methods. The essence of MIM lies in token-wise masked patch predictions, with targets patchified from images; or generated by pre-trained tokenizers or models. We argue targets from the pre-trained models usually exhibit spatial inconsistency, which makes it excessively challenging for the model to follow to learn more discriminative representations. To mitigate the issue, we introduce a novel self-supervision signal based on Dynamic Token Morphing (DTM), which dynamically aggregates contextually related tokens. DTM can be generally applied to various SSL frameworks, yet we propose a simple MIM that employs DTM to effectively improve the performance barely introducing extra training costs. Our experiments on ImageNet-1K and ADE20K evidently demonstrate the superiority of our methods. Furthermore, the comparative evaluation of iNaturalist and Fine-grained Visual Classification datasets further validates the transferability of our method on various downstream tasks. Our code will be released publicly.

Recent advances in text-to-video diffusion models have enabled high-quality video synthesis, but controllable generation remains challenging, particularly under limited data and compute. Existing fine-tuning methods for conditional generation often rely on external encoders or architectural modifications, which demand large datasets and are typically restricted to spatially aligned conditioning, limiting flexibility and scalability. In this work, we introduce Temporal In-Context Fine-Tuning (TIC-FT), an efficient and versatile approach for adapting pretrained video diffusion models to diverse conditional generation tasks. Our key idea is to concatenate condition and target frames along the temporal axis and insert intermediate buffer frames with progressively increasing noise levels. These buffer frames enable smooth transitions, aligning the fine-tuning process with the pretrained model's temporal dynamics. TIC-FT requires no architectural changes and achieves strong performance with as few as 10-30 training samples. We validate our method across a range of tasks, including image-to-video and video-to-video generation, using large-scale base models such as CogVideoX-5B and Wan-14B. Extensive experiments show that TIC-FT outperforms existing baselines in both condition fidelity and visual quality, while remaining highly efficient in both training and inference. For additional results, visit https://kinam0252.github.io/TIC-FT/

Contrastive pre-training on image-text pairs, exemplified by CLIP, becomes a standard technique for learning multi-modal visual-language representations. Although CLIP has demonstrated remarkable performance, training it from scratch on noisy web-scale datasets is computationally demanding. On the other hand, mask-then-predict pre-training approaches, like Masked Image Modeling (MIM), offer efficient self-supervised learning for single-modal representations. This paper introduces Unmasked Token Alignment (UTA), a method that leverages existing CLIP models to further enhance its vision-language representations. UTA trains a Vision Transformer (ViT) by aligning unmasked visual tokens to the corresponding image tokens from a frozen CLIP vision encoder, which automatically aligns the ViT model with the CLIP text encoder. The pre-trained ViT can be directly applied for zero-shot evaluation even without training on image-text pairs. Compared to MIM approaches, UTA does not suffer from training-finetuning inconsistency and is much more training-efficient by avoiding using the extra [MASK] tokens. Extensive experimental results demonstrate that UTA can enhance CLIP models and outperform existing MIM methods on various uni- and multi-modal benchmarks. Code and models are available at https://github.com/jihaonew/UTA.

Multi-view representation learning aims to derive robust representations that are both view-consistent and view-specific from diverse data sources. This paper presents an in-depth analysis of existing approaches in this domain, highlighting a commonly overlooked aspect: the redundancy between view-consistent and view-specific representations. To this end, we propose an innovative framework for multi-view representation learning, which incorporates a technique we term 'distilled disentangling'. Our method introduces the concept of masked cross-view prediction, enabling the extraction of compact, high-quality view-consistent representations from various sources without incurring extra computational overhead. Additionally, we develop a distilled disentangling module that efficiently filters out consistency-related information from multi-view representations, resulting in purer view-specific representations. This approach significantly reduces redundancy between view-consistent and view-specific representations, enhancing the overall efficiency of the learning process. Our empirical evaluations reveal that higher mask ratios substantially improve the quality of view-consistent representations. Moreover, we find that reducing the dimensionality of view-consistent representations relative to that of view-specific representations further refines the quality of the combined representations. Our code is accessible at: https://github.com/Guanzhou-Ke/MRDD.

Every day, humans perceive objects and communicate these perceptions through various channels. In this paper, we present a computational model designed to track and simulate the perception of objects, as well as their representations as conveyed in communication. We delineate two fundamental components of our internal representation, termed "observed" and "seen", which we correlate with established concepts in computer vision, namely encoding and decoding. These components are integrated into semiotic networks, which simulate perceptual inference of object perception and human communication. Our model of object perception by a person allows us to define object perception by {\em a network}. We demonstrate this with an example of an image baseline classifier by constructing a new network that includes the baseline classifier and an additional layer. This layer produces the images "perceived" by the entire network, transforming it into a perceptualized image classifier. This facilitates visualization of the acquired network. Within our network, the image representations become more efficient for classification tasks when they are assembled and randomized. In our experiments, the perceptualized network outperformed the baseline classifier on MNIST training databases consisting of a restricted number of images. Our model is not limited to persons and can be applied to any system featuring a loop involving the processing from "internal" to "external" representations.

This paper presents a novel approach to Zero-Shot Action Recognition. Recent works have explored the detection and classification of objects to obtain semantic information from videos with remarkable performance. Inspired by them, we propose using video captioning methods to extract semantic information about objects, scenes, humans, and their relationships. To the best of our knowledge, this is the first work to represent both videos and labels with descriptive sentences. More specifically, we represent videos using sentences generated via video captioning methods and classes using sentences extracted from documents acquired through search engines on the Internet. Using these representations, we build a shared semantic space employing BERT-based embedders pre-trained in the paraphrasing task on multiple text datasets. The projection of both visual and semantic information onto this space is straightforward, as they are sentences, enabling classification using the nearest neighbor rule. We demonstrate that representing videos and labels with sentences alleviates the domain adaptation problem. Additionally, we show that word vectors are unsuitable for building the semantic embedding space of our descriptions. Our method outperforms the state-of-the-art performance on the UCF101 dataset by 3.3 p.p. in accuracy under the TruZe protocol and achieves competitive results on both the UCF101 and HMDB51 datasets under the conventional protocol (0/50\% - training/testing split). Our code is available at https://github.com/valterlej/zsarcap.

Although Contrastive Language-Image Pre-training (CLIP) exhibits strong performance across diverse vision tasks, its application to person representation learning faces two critical challenges: (i) the scarcity of large-scale annotated vision-language data focused on person-centric images, and (ii) the inherent limitations of global contrastive learning, which struggles to maintain discriminative local features crucial for fine-grained matching while remaining vulnerable to noisy text tokens. This work advances CLIP for person representation learning through synergistic improvements in data curation and model architecture. First, we develop a noise-resistant data construction pipeline that leverages the in-context learning capabilities of MLLMs to automatically filter and caption web-sourced images. This yields WebPerson, a large-scale dataset of 5M high-quality person-centric image-text pairs. Second, we introduce the GA-DMS (Gradient-Attention Guided Dual-Masking Synergetic) framework, which improves cross-modal alignment by adaptively masking noisy textual tokens based on the gradient-attention similarity score. Additionally, we incorporate masked token prediction objectives that compel the model to predict informative text tokens, enhancing fine-grained semantic representation learning. Extensive experiments show that GA-DMS achieves state-of-the-art performance across multiple benchmarks.

We present a frame interpolation algorithm that synthesizes multiple intermediate frames from two input images with large in-between motion. Recent methods use multiple networks to estimate optical flow or depth and a separate network dedicated to frame synthesis. This is often complex and requires scarce optical flow or depth ground-truth. In this work, we present a single unified network, distinguished by a multi-scale feature extractor that shares weights at all scales, and is trainable from frames alone. To synthesize crisp and pleasing frames, we propose to optimize our network with the Gram matrix loss that measures the correlation difference between feature maps. Our approach outperforms state-of-the-art methods on the Xiph large motion benchmark. We also achieve higher scores on Vimeo-90K, Middlebury and UCF101, when comparing to methods that use perceptual losses. We study the effect of weight sharing and of training with datasets of increasing motion range. Finally, we demonstrate our model's effectiveness in synthesizing high quality and temporally coherent videos on a challenging near-duplicate photos dataset. Codes and pre-trained models are available at https://film-net.github.io.

Large Vision-Language Models (LVLMs) have demonstrated remarkable performance across various multimodal tasks by integrating visual perception with language understanding. However, conventional decoding strategies of LVLMs often fail to successfully utilize visual information, leading to visually ungrounded responses. While various approaches have been proposed to address this limitation, they typically require additional training, multi-step inference procedures, or external model dependencies. This paper introduces ReVisiT, a simple yet effective decoding method that references vision tokens to guide the text generation process in LVLMs. Our approach leverages the semantic information embedded within vision tokens by projecting them into the text token distribution space, and dynamically selecting the most relevant vision token at each decoding step through constrained divergence minimization. This selected vision token is then used to refine the output distribution to better incorporate visual semantics. Experiments on three LVLM hallucination benchmarks with two recent LVLMs demonstrate that ReVisiT consistently enhances visual grounding with minimal computational overhead. Moreover, our method achieves competitive or superior results relative to state-of-the-art baselines while reducing computational costs for up to 2times.

Vision-and-language reasoning requires an understanding of visual concepts, language semantics, and, most importantly, the alignment and relationships between these two modalities. We thus propose the LXMERT (Learning Cross-Modality Encoder Representations from Transformers) framework to learn these vision-and-language connections. In LXMERT, we build a large-scale Transformer model that consists of three encoders: an object relationship encoder, a language encoder, and a cross-modality encoder. Next, to endow our model with the capability of connecting vision and language semantics, we pre-train the model with large amounts of image-and-sentence pairs, via five diverse representative pre-training tasks: masked language modeling, masked object prediction (feature regression and label classification), cross-modality matching, and image question answering. These tasks help in learning both intra-modality and cross-modality relationships. After fine-tuning from our pre-trained parameters, our model achieves the state-of-the-art results on two visual question answering datasets (i.e., VQA and GQA). We also show the generalizability of our pre-trained cross-modality model by adapting it to a challenging visual-reasoning task, NLVR2, and improve the previous best result by 22% absolute (54% to 76%). Lastly, we demonstrate detailed ablation studies to prove that both our novel model components and pre-training strategies significantly contribute to our strong results; and also present several attention visualizations for the different encoders. Code and pre-trained models publicly available at: https://github.com/airsplay/lxmert

Recent Video-Language Models (VLMs) achieve promising results on long-video understanding, but their performance still lags behind that achieved on tasks involving images or short videos. This has led to great interest in improving the long context modeling of VLMs by introducing novel modules and additional complexity. % additional training time. In this paper, we take a different approach: rather than fine-tuning VLMs with the limited data available, we attempt to maximize the performance of existing models. To this end, we propose a novel visual prompting strategy specifically designed for long-video understanding. By combining multiple frames as panels into one image, we effectively trade off spatial details for temporal resolution. Our approach is training-free, parameter-free, and model-agnostic, and can be seamlessly integrated into existing VLMs. Extensive experiments on five established benchmarks across a wide range of model architectures, sizes, and context windows confirm the consistency of our approach. For the TimeScope (Long) dataset, which has the longest videos, the accuracy for video question answering is improved by up to 19.4\%. Overall, our method raises the bar for long video understanding models. We will make our code available upon acceptance.

There has been a longstanding belief that generation can facilitate a true understanding of visual data. In line with this, we revisit generatively pre-training visual representations in light of recent interest in denoising diffusion models. While directly pre-training with diffusion models does not produce strong representations, we condition diffusion models on masked input and formulate diffusion models as masked autoencoders (DiffMAE). Our approach is capable of (i) serving as a strong initialization for downstream recognition tasks, (ii) conducting high-quality image inpainting, and (iii) being effortlessly extended to video where it produces state-of-the-art classification accuracy. We further perform a comprehensive study on the pros and cons of design choices and build connections between diffusion models and masked autoencoders.

In this paper, we present an integral pre-training framework based on masked image modeling (MIM). We advocate for pre-training the backbone and neck jointly so that the transfer gap between MIM and downstream recognition tasks is minimal. We make two technical contributions. First, we unify the reconstruction and recognition necks by inserting a feature pyramid into the pre-training stage. Second, we complement mask image modeling (MIM) with masked feature modeling (MFM) that offers multi-stage supervision to the feature pyramid. The pre-trained models, termed integrally pre-trained transformer pyramid networks (iTPNs), serve as powerful foundation models for visual recognition. In particular, the base/large-level iTPN achieves an 86.2%/87.8% top-1 accuracy on ImageNet-1K, a 53.2%/55.6% box AP on COCO object detection with 1x training schedule using Mask-RCNN, and a 54.7%/57.7% mIoU on ADE20K semantic segmentation using UPerHead -- all these results set new records. Our work inspires the community to work on unifying upstream pre-training and downstream fine-tuning tasks. Code and the pre-trained models will be released at https://github.com/sunsmarterjie/iTPN.

Efficient understanding of long-form videos remains a significant challenge in computer vision. In this work, we revisit temporal search paradigms for long-form video understanding, studying a fundamental issue pertaining to all state-of-the-art (SOTA) long-context vision-language models (VLMs). In particular, our contributions are two-fold: First, we formulate temporal search as a Long Video Haystack problem, i.e., finding a minimal set of relevant frames (typically one to five) among tens of thousands of frames from real-world long videos given specific queries. To validate our formulation, we create LV-Haystack, the first benchmark containing 3,874 human-annotated instances with fine-grained evaluation metrics for assessing keyframe search quality and computational efficiency. Experimental results on LV-Haystack highlight a significant research gap in temporal search capabilities, with SOTA keyframe selection methods achieving only 2.1% temporal F1 score on the LVBench subset. Next, inspired by visual search in images, we re-think temporal searching and propose a lightweight keyframe searching framework, T*, which casts the expensive temporal search as a spatial search problem. T* leverages superior visual localization capabilities typically used in images and introduces an adaptive zooming-in mechanism that operates across both temporal and spatial dimensions. Our extensive experiments show that when integrated with existing methods, T* significantly improves SOTA long-form video understanding performance. Specifically, under an inference budget of 32 frames, T* improves GPT-4o's performance from 50.5% to 53.1% and LLaVA-OneVision-72B's performance from 56.5% to 62.4% on LongVideoBench XL subset. Our PyTorch code, benchmark dataset and models are included in the Supplementary material.

As deep vision models' popularity rapidly increases, there is a growing emphasis on explanations for model predictions. The inherently explainable attribution method aims to enhance the understanding of model behavior by identifying the important regions in images that significantly contribute to predictions. It is achieved by cooperatively training a selector (generating an attribution map to identify important features) and a predictor (making predictions using the identified features). Despite many advancements, existing methods suffer from the incompleteness problem, where discriminative features are masked out, and the interlocking problem, where the non-optimized selector initially selects noise, causing the predictor to fit on this noise and perpetuate the cycle. To address these problems, we introduce a new objective that discourages the presence of discriminative features in the masked-out regions thus enhancing the comprehensiveness of feature selection. A pre-trained detector is introduced to detect discriminative features in the masked-out region. If the selector selects noise instead of discriminative features, the detector can observe and break the interlocking situation by penalizing the selector. Extensive experiments show that our model makes accurate predictions with higher accuracy than the regular black-box model, and produces attribution maps with high feature coverage, localization ability, fidelity and robustness. Our code will be available at https://github.com/Zood123/COMET{https://github.com/Zood123/COMET}.

Multimodal Large Language Models (MLLMs) have significantly improved performance across various image-language applications. Recently, there has been a growing interest in adapting image pre-trained MLLMs for video-related tasks. However, most efforts concentrate on enhancing the vision encoder and projector components, while the core part, Large Language Models (LLMs), remains comparatively under-explored. In this paper, we propose two strategies to enhance the model's capability in video understanding tasks by improving inter-layer attention computation in LLMs. Specifically, the first approach focuses on the enhancement of Rotary Position Embedding (RoPE) with Temporal-Aware Dual RoPE, which introduces temporal position information to strengthen the MLLM's temporal modeling capabilities while preserving the relative position relationships of both visual and text tokens. The second approach involves enhancing the Attention Mask with the Frame-wise Block Causal Attention Mask, a simple yet effective method that broadens visual token interactions within and across video frames while maintaining the causal inference mechanism. Based on these proposed methods, we adapt LLaVA for video understanding tasks, naming it Temporal-Considered LLaVA (TC-LLaVA). Our TC-LLaVA achieves new state-of-the-art performance across various video understanding benchmarks with only supervised fine-tuning (SFT) on video-related datasets.

We propose a pre-training strategy called Multi-modal Multi-task Masked Autoencoders (MultiMAE). It differs from standard Masked Autoencoding in two key aspects: I) it can optionally accept additional modalities of information in the input besides the RGB image (hence "multi-modal"), and II) its training objective accordingly includes predicting multiple outputs besides the RGB image (hence "multi-task"). We make use of masking (across image patches and input modalities) to make training MultiMAE tractable as well as to ensure cross-modality predictive coding is indeed learned by the network. We show this pre-training strategy leads to a flexible, simple, and efficient framework with improved transfer results to downstream tasks. In particular, the same exact pre-trained network can be flexibly used when additional information besides RGB images is available or when no information other than RGB is available - in all configurations yielding competitive to or significantly better results than the baselines. To avoid needing training datasets with multiple modalities and tasks, we train MultiMAE entirely using pseudo labeling, which makes the framework widely applicable to any RGB dataset. The experiments are performed on multiple transfer tasks (image classification, semantic segmentation, depth estimation) and datasets (ImageNet, ADE20K, Taskonomy, Hypersim, NYUv2). The results show an intriguingly impressive capability by the model in cross-modal/task predictive coding and transfer.

Multimodal Large Language Models (MLLMs) have shown exceptional capabilities in vision-language tasks. However, effectively integrating image segmentation into these models remains a significant challenge. In this work, we propose a novel text-as-mask paradigm that casts image segmentation as a text generation problem, eliminating the need for additional decoders and significantly simplifying the segmentation process. Our key innovation is semantic descriptors, a new textual representation of segmentation masks where each image patch is mapped to its corresponding text label. We first introduce image-wise semantic descriptors, a patch-aligned textual representation of segmentation masks that integrates naturally into the language modeling pipeline. To enhance efficiency, we introduce the Row-wise Run-Length Encoding (R-RLE), which compresses redundant text sequences, reducing the length of semantic descriptors by 74% and accelerating inference by 3times, without compromising performance. Building upon this, our initial framework Text4Seg achieves strong segmentation performance across a wide range of vision tasks. To further improve granularity and compactness, we propose box-wise semantic descriptors, which localizes regions of interest using bounding boxes and represents region masks via structured mask tokens called semantic bricks. This leads to our refined model, Text4Seg++, which formulates segmentation as a next-brick prediction task, combining precision, scalability, and generative efficiency. Comprehensive experiments on natural and remote sensing datasets show that Text4Seg++ consistently outperforms state-of-the-art models across diverse benchmarks without any task-specific fine-tuning, while remaining compatible with existing MLLM backbones. Our work highlights the effectiveness, scalability, and generalizability of text-driven image segmentation within the MLLM framework.

The best summary of a long video differs among different people due to its highly subjective nature. Even for the same person, the best summary may change with time or mood. In this paper, we introduce the task of generating customized video summaries through simple text. First, we train a deep architecture to effectively learn semantic embeddings of video frames by leveraging the abundance of image-caption data via a progressive and residual manner. Given a user-specific text description, our algorithm is able to select semantically relevant video segments and produce a temporally aligned video summary. In order to evaluate our textually customized video summaries, we conduct experimental comparison with baseline methods that utilize ground-truth information. Despite the challenging baselines, our method still manages to show comparable or even exceeding performance. We also show that our method is able to generate semantically diverse video summaries by only utilizing the learned visual embeddings.

We address the problem of detecting human-object interactions in images using graphical neural networks. Unlike conventional methods, where nodes send scaled but otherwise identical messages to each of their neighbours, we propose to condition messages between pairs of nodes on their spatial relationships, resulting in different messages going to neighbours of the same node. To this end, we explore various ways of applying spatial conditioning under a multi-branch structure. Through extensive experimentation we demonstrate the advantages of spatial conditioning for the computation of the adjacency structure, messages and the refined graph features. In particular, we empirically show that as the quality of the bounding boxes increases, their coarse appearance features contribute relatively less to the disambiguation of interactions compared to the spatial information. Our method achieves an mAP of 31.33% on HICO-DET and 54.2% on V-COCO, significantly outperforming state-of-the-art on fine-tuned detections.

Masked Autoencoders (MAE) have been prevailing paradigms for large-scale vision representation pre-training. By reconstructing masked image patches from a small portion of visible image regions, MAE forces the model to infer semantic correlation within an image. Recently, some approaches apply semantic-rich teacher models to extract image features as the reconstruction target, leading to better performance. However, unlike the low-level features such as pixel values, we argue the features extracted by powerful teacher models already encode rich semantic correlation across regions in an intact image.This raises one question: is reconstruction necessary in Masked Image Modeling (MIM) with a teacher model? In this paper, we propose an efficient MIM paradigm named MaskAlign. MaskAlign simply learns the consistency of visible patch features extracted by the student model and intact image features extracted by the teacher model. To further advance the performance and tackle the problem of input inconsistency between the student and teacher model, we propose a Dynamic Alignment (DA) module to apply learnable alignment. Our experimental results demonstrate that masked modeling does not lose effectiveness even without reconstruction on masked regions. Combined with Dynamic Alignment, MaskAlign can achieve state-of-the-art performance with much higher efficiency. Code and models will be available at https://github.com/OpenPerceptionX/maskalign.

We introduce an object-aware decoder for improving the performance of spatio-temporal representations on ego-centric videos. The key idea is to enhance object-awareness during training by tasking the model to predict hand positions, object positions, and the semantic label of the objects using paired captions when available. At inference time the model only requires RGB frames as inputs, and is able to track and ground objects (although it has not been trained explicitly for this). We demonstrate the performance of the object-aware representations learnt by our model, by: (i) evaluating it for strong transfer, i.e. through zero-shot testing, on a number of downstream video-text retrieval and classification benchmarks; and (ii) by using the representations learned as input for long-term video understanding tasks (e.g. Episodic Memory in Ego4D). In all cases the performance improves over the state of the art -- even compared to networks trained with far larger batch sizes. We also show that by using noisy image-level detection as pseudo-labels in training, the model learns to provide better bounding boxes using video consistency, as well as grounding the words in the associated text descriptions. Overall, we show that the model can act as a drop-in replacement for an ego-centric video model to improve performance through visual-text grounding.

Video reasoning segmentation (VRS) endeavors to delineate referred objects in videos guided by implicit instructions that encapsulate human intent and temporal logic. Previous approaches leverage large vision language models (LVLMs) to encode object semantics into <SEG> tokens for mask prediction. However, this paradigm suffers from limited interpretability during inference and suboptimal performance due to inadequate spatiotemporal reasoning. Drawing inspiration from seminal breakthroughs in reinforcement learning, we introduce Veason-R1, a specialized LVLM for VRS that emphasizes structured reasoning in segmentation. Veason-R1 is trained through Group Relative Policy Optimization (GRPO) augmented with Chain-of-Thought (CoT) initialization. To begin with, we curate high-quality CoT training data to instill structured reasoning trajectories, bridging video-level semantics and frame-level spatial grounding, yielding the supervised fine-tuned model Veason-SFT. Subsequently, GRPO fine-tuning encourages efficient exploration of the reasoning space by optimizing reasoning chains. To this end, we incorporate a holistic reward mechanism that synergistically enhances spatial alignment and temporal consistency, bolstering keyframe localization and fine-grained grounding. Comprehensive empirical evaluations demonstrate that Veason-R1 achieves state-of-the-art performance on multiple benchmarks, surpassing prior art by significant margins (e.g., +1.3 J &F in ReVOS and +10.0 J &F in ReasonVOS), while exhibiting robustness to hallucinations (+8.8 R). Our code and model weights will be available at Veason-R1.

To build scalable models for challenging real-world tasks, it is important to learn from diverse, multi-modal data in various forms (e.g., videos, text, and images). Among the existing works, a plethora of them have focused on leveraging large but cumbersome cross-modal architectures. Regardless of their effectiveness, larger architectures unavoidably prevent the models from being extended to real-world applications, so building a lightweight VL architecture and an efficient learning schema is of great practical value. In this paper, we propose an Efficient Video-Language Model (dubbed as E-ViLM) and a masked video modeling (MVM) schema, assisted with a semantic vector-quantized tokenizer. In particular, our E-ViLM learns to reconstruct the semantic labels of masked video regions, produced by the pre-trained vector-quantized tokenizer, which discretizes the continuous visual signals into labels. We show that with our simple MVM task and regular VL pre-training modelings, our E-ViLM, despite its compactness, is able to learn expressive representations from Video-Language corpus and generalize well to extensive Video-Language tasks including video question answering, text-to-video retrieval, etc. In particular, our E-ViLM obtains obvious efficiency improvements by reaching competing performances with faster inference speed, i.e., our model reaches 39.3% Top-1 accuracy on the MSRVTT benchmark, retaining 91.4% of the accuracy of state-of-the-art larger VL architecture with only 15% parameters and 94.8% fewer GFLOPs. We also provide extensive ablative studies that validate the effectiveness of our proposed learning schema for E-ViLM.

Large Multimodal Models have demonstrated impressive capabilities in understanding general vision-language tasks. However, due to the limitation of supported input resolution (e.g., 448 x 448) as well as the inexhaustive description of the training image-text pair, these models often encounter challenges when dealing with intricate scene understandings and narratives. Here we address the problem by proposing the Monkey. Our contributions are two-fold: 1) without pretraining from the start, our method can be built upon an existing vision encoder (e.g., vit-BigHuge) to effectively improve the input resolution capacity up to 896 x 1344 pixels; 2) we propose a multi-level description generation method, which automatically provides rich information that can guide model to learn contextual association between scenes and objects. Our extensive testing across more than 16 distinct datasets reveals that Monkey achieves consistently competitive performance over the existing LMMs on fundamental tasks, such as Image Captioning, General Visual Question Answering (VQA), and Document-oriented VQA. Models, interactive demo, and the source code are provided at the following https://github.com/Yuliang-Liu/Monkey.

While recent progress in video-text retrieval has been advanced by the exploration of better representation learning, in this paper, we present a novel multi-grained sparse learning framework, S3MA, to learn an aligned sparse space shared between the video and the text for video-text retrieval. The shared sparse space is initialized with a finite number of sparse concepts, each of which refers to a number of words. With the text data at hand, we learn and update the shared sparse space in a supervised manner using the proposed similarity and alignment losses. Moreover, to enable multi-grained alignment, we incorporate frame representations for better modeling the video modality and calculating fine-grained and coarse-grained similarities. Benefiting from the learned shared sparse space and multi-grained similarities, extensive experiments on several video-text retrieval benchmarks demonstrate the superiority of S3MA over existing methods. Our code is available at https://github.com/yimuwangcs/Better_Cross_Modal_Retrieval.

Large-scale pretrained image-text models have shown incredible zero-shot performance in a handful of tasks, including video ones such as action recognition and text-to-video retrieval. However, these models have not been adapted to video, mainly because they do not account for the time dimension but also because video frames are different from the typical images (e.g., containing motion blur, and less sharpness). In this paper, we present a fine-tuning strategy to refine these large-scale pretrained image-text models for zero-shot video understanding tasks. We show that by carefully adapting these models we obtain considerable improvements on two zero-shot Action Recognition tasks and three zero-shot Text-to-video Retrieval tasks. The code is available at https://github.com/bryant1410/fitclip

We propose a novel feed-forward network for video inpainting. We use a set of sampled video frames as the reference to take visible contents to fill the hole of a target frame. Our video inpainting network consists of two stages. The first stage is an alignment module that uses computed homographies between the reference frames and the target frame. The visible patches are then aggregated based on the frame similarity to fill in the target holes roughly. The second stage is a non-local attention module that matches the generated patches with known reference patches (in space and time) to refine the previous global alignment stage. Both stages consist of large spatial-temporal window size for the reference and thus enable modeling long-range correlations between distant information and the hole regions. Therefore, even challenging scenes with large or slowly moving holes can be handled, which have been hardly modeled by existing flow-based approach. Our network is also designed with a recurrent propagation stream to encourage temporal consistency in video results. Experiments on video object removal demonstrate that our method inpaints the holes with globally and locally coherent contents.

In this paper, we introduce the task of automatically generating text to describe the differences between two similar images. We collect a new dataset by crowd-sourcing difference descriptions for pairs of image frames extracted from video-surveillance footage. Annotators were asked to succinctly describe all the differences in a short paragraph. As a result, our novel dataset provides an opportunity to explore models that align language and vision, and capture visual salience. The dataset may also be a useful benchmark for coherent multi-sentence generation. We perform a firstpass visual analysis that exposes clusters of differing pixels as a proxy for object-level differences. We propose a model that captures visual salience by using a latent variable to align clusters of differing pixels with output sentences. We find that, for both single-sentence generation and as well as multi-sentence generation, the proposed model outperforms the models that use attention alone.

In this work, we introduce long-video masked-embedding autoencoders (LV-MAE), a self-supervised learning framework for long video representation. Our approach treats short- and long-span dependencies as two separate tasks. Such decoupling allows for a more intuitive video processing where short-span spatiotemporal primitives are first encoded and are then used to capture long-range dependencies across consecutive video segments. To achieve this, we leverage advanced off-the-shelf multimodal encoders to extract representations from short segments within the long video, followed by pre-training a masked-embedding autoencoder capturing high-level interactions across segments. LV-MAE is highly efficient to train and enables the processing of much longer videos by alleviating the constraint on the number of input frames. Furthermore, unlike existing methods that typically pre-train on short-video datasets, our approach offers self-supervised pre-training using long video samples (e.g., 20+ minutes video clips) at scale. Using LV-MAE representations, we achieve state-of-the-art results on three long-video benchmarks -- LVU, COIN, and Breakfast -- employing only a simple classification head for either attentive or linear probing. Finally, to assess LV-MAE pre-training and visualize its reconstruction quality, we leverage the video-language aligned space of short video representations to monitor LV-MAE through video-text retrieval.

Masked image modeling (MIM) has become a leading self-supervised learning strategy. MIMs such as Masked Autoencoder (MAE) learn strong representations by randomly masking input tokens for the encoder to process, with the decoder reconstructing the masked tokens to the input. However, MIM pre-trained encoders often exhibit a limited attention span, attributed to MIM's sole focus on regressing masked tokens only, which may impede the encoder's broader context learning. To tackle the limitation, we improve MIM by explicitly incorporating unmasked tokens into the training process. Specifically, our method enables the encoder to learn from broader context supervision, allowing unmasked tokens to experience broader contexts while the decoder reconstructs masked tokens. Thus, the encoded unmasked tokens are equipped with extensive contextual information, empowering masked tokens to leverage the enhanced unmasked tokens for MIM. As a result, our simple remedy trains more discriminative representations revealed by achieving 84.2% top-1 accuracy with ViT-B on ImageNet-1K with 0.6%p gain. We attribute the success to the enhanced pre-training method, as evidenced by the singular value spectrum and attention analyses. Finally, our models achieve significant performance gains at the downstream semantic segmentation and fine-grained visual classification tasks; and on diverse robust evaluation metrics. Code is available at https://github.com/naver-ai/lut

Masked video modeling, such as VideoMAE, is an effective paradigm for video self-supervised learning (SSL). However, they are primarily based on reconstructing pixel-level details on natural videos which have substantial temporal redundancy, limiting their capability for semantic representation and sufficient encoding of motion dynamics. To address these issues, this paper introduces a novel SSL approach for video representation learning, dubbed as SMILE, by infusing both spatial and motion semantics. In SMILE, we leverage image-language pretrained models, such as CLIP, to guide the learning process with their high-level spatial semantics. We enhance the representation of motion by introducing synthetic motion patterns in the training data, allowing the model to capture more complex and dynamic content. Furthermore, using SMILE, we establish a new self-supervised video learning paradigm capable of learning strong video representations without requiring any natural video data. We have carried out extensive experiments on 7 datasets with various downstream scenarios. SMILE surpasses current state-of-the-art SSL methods, showcasing its effectiveness in learning more discriminative and generalizable video representations. Code is available: https://github.com/fmthoker/SMILE

Diffusion models have shown remarkable progress in various generative tasks such as image and video generation. This paper studies the problem of leveraging pretrained diffusion models for performing discriminative tasks. Specifically, we extend the discriminative capability of pretrained frozen generative diffusion models from the classification task to the more complex object detection task, by "inverting" a pretrained layout-to-image diffusion model. To this end, a gradient-based discrete optimization approach for replacing the heavy prediction enumeration process, and a prior distribution model for making more accurate use of the Bayes' rule, are proposed respectively. Empirical results show that this method is on par with basic discriminative object detection baselines on COCO dataset. In addition, our method can greatly speed up the previous diffusion-based method for classification without sacrificing accuracy. Code and models are available at https://github.com/LiYinqi/DIVE .

Fine-tuning Video Diffusion Models (VDMs) at the user level to generate videos that reflect specific attributes of training data presents notable challenges, yet remains underexplored despite its practical importance. Meanwhile, recent work such as Representation Alignment (REPA) has shown promise in improving the convergence and quality of DiT-based image diffusion models by aligning, or assimilating, its internal hidden states with external pretrained visual features, suggesting its potential for VDM fine-tuning. In this work, we first propose a straightforward adaptation of REPA for VDMs and empirically show that, while effective for convergence, it is suboptimal in preserving semantic consistency across frames. To address this limitation, we introduce Cross-frame Representation Alignment (CREPA), a novel regularization technique that aligns hidden states of a frame with external features from neighboring frames. Empirical evaluations on large-scale VDMs, including CogVideoX-5B and Hunyuan Video, demonstrate that CREPA improves both visual fidelity and cross-frame semantic coherence when fine-tuned with parameter-efficient methods such as LoRA. We further validate CREPA across diverse datasets with varying attributes, confirming its broad applicability. Project page: https://crepavideo.github.io

This paper presents SimMIM, a simple framework for masked image modeling. We simplify recently proposed related approaches without special designs such as block-wise masking and tokenization via discrete VAE or clustering. To study what let the masked image modeling task learn good representations, we systematically study the major components in our framework, and find that simple designs of each component have revealed very strong representation learning performance: 1) random masking of the input image with a moderately large masked patch size (e.g., 32) makes a strong pre-text task; 2) predicting raw pixels of RGB values by direct regression performs no worse than the patch classification approaches with complex designs; 3) the prediction head can be as light as a linear layer, with no worse performance than heavier ones. Using ViT-B, our approach achieves 83.8% top-1 fine-tuning accuracy on ImageNet-1K by pre-training also on this dataset, surpassing previous best approach by +0.6%. When applied on a larger model of about 650 million parameters, SwinV2-H, it achieves 87.1% top-1 accuracy on ImageNet-1K using only ImageNet-1K data. We also leverage this approach to facilitate the training of a 3B model (SwinV2-G), that by 40times less data than that in previous practice, we achieve the state-of-the-art on four representative vision benchmarks. The code and models will be publicly available at https://github.com/microsoft/SimMIM.

Video multimodal large language models (Video-MLLMs) have achieved remarkable progress in video understanding. However, they remain vulnerable to hallucination-producing content inconsistent with or unrelated to video inputs. Previous video hallucination benchmarks primarily focus on short-videos. They attribute hallucinations to factors such as strong language priors, missing frames, or vision-language biases introduced by the visual encoder. While these causes indeed account for most hallucinations in short videos, they still oversimplify the cause of hallucinations. Sometimes, models generate incorrect outputs but with correct frame-level semantics. We refer to this type of hallucination as Semantic Aggregation Hallucination (SAH), which arises during the process of aggregating frame-level semantics into event-level semantic groups. Given that SAH becomes particularly critical in long videos due to increased semantic complexity across multiple events, it is essential to separate and thoroughly investigate the causes of this type of hallucination. To address the above issues, we introduce ELV-Halluc, the first benchmark dedicated to long-video hallucination, enabling a systematic investigation of SAH. Our experiments confirm the existence of SAH and show that it increases with semantic complexity. Additionally, we find that models are more prone to SAH on rapidly changing semantics. Moreover, we discuss potential approaches to mitigate SAH. We demonstrate that positional encoding strategy contributes to alleviating SAH, and further adopt DPO strategy to enhance the model's ability to distinguish semantics within and across events. To support this, we curate a dataset of 8K adversarial data pairs and achieve improvements on both ELV-Halluc and Video-MME, including a substantial 27.7% reduction in SAH ratio.

We present Compound Conditioned ControlNet, C3Net, a novel generative neural architecture taking conditions from multiple modalities and synthesizing multimodal contents simultaneously (e.g., image, text, audio). C3Net adapts the ControlNet architecture to jointly train and make inferences on a production-ready diffusion model and its trainable copies. Specifically, C3Net first aligns the conditions from multi-modalities to the same semantic latent space using modality-specific encoders based on contrastive training. Then, it generates multimodal outputs based on the aligned latent space, whose semantic information is combined using a ControlNet-like architecture called Control C3-UNet. Correspondingly, with this system design, our model offers an improved solution for joint-modality generation through learning and explaining multimodal conditions instead of simply taking linear interpolations on the latent space. Meanwhile, as we align conditions to a unified latent space, C3Net only requires one trainable Control C3-UNet to work on multimodal semantic information. Furthermore, our model employs unimodal pretraining on the condition alignment stage, outperforming the non-pretrained alignment even on relatively scarce training data and thus demonstrating high-quality compound condition generation. We contribute the first high-quality tri-modal validation set to validate quantitatively that C3Net outperforms or is on par with first and contemporary state-of-the-art multimodal generation. Our codes and tri-modal dataset will be released.

Masked image modeling (MIM) has attracted much research attention due to its promising potential for learning scalable visual representations. In typical approaches, models usually focus on predicting specific contents of masked patches, and their performances are highly related to pre-defined mask strategies. Intuitively, this procedure can be considered as training a student (the model) on solving given problems (predict masked patches). However, we argue that the model should not only focus on solving given problems, but also stand in the shoes of a teacher to produce a more challenging problem by itself. To this end, we propose Hard Patches Mining (HPM), a brand-new framework for MIM pre-training. We observe that the reconstruction loss can naturally be the metric of the difficulty of the pre-training task. Therefore, we introduce an auxiliary loss predictor, predicting patch-wise losses first and deciding where to mask next. It adopts a relative relationship learning strategy to prevent overfitting to exact reconstruction loss values. Experiments under various settings demonstrate the effectiveness of HPM in constructing masked images. Furthermore, we empirically find that solely introducing the loss prediction objective leads to powerful representations, verifying the efficacy of the ability to be aware of where is hard to reconstruct.

Despite its success in image synthesis, we observe that diffusion probabilistic models (DPMs) often lack contextual reasoning ability to learn the relations among object parts in an image, leading to a slow learning process. To solve this issue, we propose a Masked Diffusion Transformer (MDT) that introduces a mask latent modeling scheme to explicitly enhance the DPMs' ability of contextual relation learning among object semantic parts in an image. During training, MDT operates on the latent space to mask certain tokens. Then, an asymmetric masking diffusion transformer is designed to predict masked tokens from unmasked ones while maintaining the diffusion generation process. Our MDT can reconstruct the full information of an image from its incomplete contextual input, thus enabling it to learn the associated relations among image tokens. Experimental results show that MDT achieves superior image synthesis performance, e.g. a new SoTA FID score on the ImageNet dataset, and has about 3x faster learning speed than the previous SoTA DiT. The source code is released at https://github.com/sail-sg/MDT.

Contrastive language-image pretraining (CLIP) has significantly advanced image-based vision learning. A pressing topic subsequently arises: how can we effectively adapt CLIP to the video domain? Recent studies have focused on adjusting either the textual or visual branch of CLIP for action recognition. However, we argue that adaptations of both branches are crucial. In this paper, we propose CLAVER: a Contrastive Language-Action Video Learner, designed to shift CLIP's focus from the alignment of static visual objects and concrete nouns to the alignment of dynamic action behaviors and abstract verbs. Specifically, we introduce a novel Kronecker mask attention for temporal modeling. Our tailored Kronecker mask offers three benefits 1) it expands the temporal receptive field for each token, 2) it serves as an effective spatiotemporal heterogeneity inductive bias, mitigating the issue of spatiotemporal homogenization, and 3) it can be seamlessly plugged into transformer-based models. Regarding the textual branch, we leverage large language models to generate diverse, sentence-level and semantically rich interpretive prompts of actions, which shift the model's focus towards the verb comprehension. Extensive experiments on various benchmarks and learning scenarios demonstrate the superiority and generality of our approach.

Building video-language foundation models is costly and difficult due to the redundant nature of video data and the lack of high-quality video-language datasets. In this paper, we propose an efficient framework to harvest video foundation models from image ones. Our method is intuitively simple by randomly dropping input video patches and masking out input text during the post-pretraining procedure. The patch dropping boosts the training efficiency significantly and text masking enforces the learning of cross-modal fusion. We conduct extensive experiments to validate the effectiveness of our method on a wide range of video-language downstream tasks including various zero-shot tasks, video question answering, and video-text retrieval. Despite its simplicity, our method achieves state-of-the-art performances, which are comparable to some heavily pretrained video foundation models. Our method is extremely efficient and can be trained in less than one day on 8 GPUs, requiring only WebVid-10M as pretraining data. We hope our method can serve as a simple yet strong counterpart for prevalent video foundation models, provide useful insights when building them, and make large pretrained models more accessible and sustainable. This is part of the InternVideo project https://github.com/OpenGVLab/InternVideo.

Video-based pretraining offers immense potential for learning strong visual representations on an unprecedented scale. Recently, masked video modeling methods have shown promising scalability, yet fall short in capturing higher-level semantics due to reconstructing predefined low-level targets such as pixels. To tackle this, we present Sinkhorn-guided Masked Video Modelling (SIGMA), a novel video pretraining method that jointly learns the video model in addition to a target feature space using a projection network. However, this simple modification means that the regular L2 reconstruction loss will lead to trivial solutions as both networks are jointly optimized. As a solution, we distribute features of space-time tubes evenly across a limited number of learnable clusters. By posing this as an optimal transport problem, we enforce high entropy in the generated features across the batch, infusing semantic and temporal meaning into the feature space. The resulting cluster assignments are used as targets for a symmetric prediction task where the video model predicts cluster assignment of the projection network and vice versa. Experimental results on ten datasets across three benchmarks validate the effectiveness of SIGMA in learning more performant, temporally-aware, and robust video representations improving upon state-of-the-art methods. Our project website with code is available at: https://quva-lab.github.io/SIGMA.

We propose a method for jointly inferring labels across a collection of data samples, where each sample consists of an observation and a prior belief about the label. By implicitly assuming the existence of a generative model for which a differentiable predictor is the posterior, we derive a training objective that allows learning under weak beliefs. This formulation unifies various machine learning settings; the weak beliefs can come in the form of noisy or incomplete labels, likelihoods given by a different prediction mechanism on auxiliary input, or common-sense priors reflecting knowledge about the structure of the problem at hand. We demonstrate the proposed algorithms on diverse problems: classification with negative training examples, learning from rankings, weakly and self-supervised aerial imagery segmentation, co-segmentation of video frames, and coarsely supervised text classification.

Video generation has made substantial strides with the emergence of deep generative models, especially diffusion-based approaches. However, video generation based on multiple reference subjects still faces significant challenges in maintaining multi-subject consistency and ensuring high generation quality. In this paper, we propose MAGREF, a unified framework for any-reference video generation that introduces masked guidance to enable coherent multi-subject video synthesis conditioned on diverse reference images and a textual prompt. Specifically, we propose (1) a region-aware dynamic masking mechanism that enables a single model to flexibly handle various subject inference, including humans, objects, and backgrounds, without architectural changes, and (2) a pixel-wise channel concatenation mechanism that operates on the channel dimension to better preserve appearance features. Our model delivers state-of-the-art video generation quality, generalizing from single-subject training to complex multi-subject scenarios with coherent synthesis and precise control over individual subjects, outperforming existing open-source and commercial baselines. To facilitate evaluation, we also introduce a comprehensive multi-subject video benchmark. Extensive experiments demonstrate the effectiveness of our approach, paving the way for scalable, controllable, and high-fidelity multi-subject video synthesis. Code and model can be found at: https://github.com/MAGREF-Video/MAGREF

Recent advances in vision-language models have significantly propelled video understanding. Existing datasets and tasks, however, have notable limitations. Most datasets are confined to short videos with limited events and narrow narratives. For example, datasets with instructional and egocentric videos often document the activities of one person in a single scene. Although some movie datasets offer richer content, they are often limited to short-term tasks, lack publicly available videos and frequently encounter data leakage given the use of movie forums and other resources in LLM training. To address the above limitations, we propose the Short Film Dataset (SFD) with 1,078 publicly available amateur movies, a wide variety of genres and minimal data leakage issues. SFD offers long-term story-oriented video tasks in the form of multiple-choice and open-ended question answering. Our extensive experiments emphasize the need for long-term reasoning to solve SFD tasks. Notably, we find strong signals in movie transcripts leading to the on-par performance of people and LLMs. We also show significantly lower performance of current models compared to people when using vision data alone.

Inspired by the remarkable success of Latent Diffusion Models (LDMs) for image synthesis, we study LDM for text-to-video generation, which is a formidable challenge due to the computational and memory constraints during both model training and inference. A single LDM is usually only capable of generating a very limited number of video frames. Some existing works focus on separate prediction models for generating more video frames, which suffer from additional training cost and frame-level jittering, however. In this paper, we propose a framework called "Reuse and Diffuse" dubbed VidRD to produce more frames following the frames already generated by an LDM. Conditioned on an initial video clip with a small number of frames, additional frames are iteratively generated by reusing the original latent features and following the previous diffusion process. Besides, for the autoencoder used for translation between pixel space and latent space, we inject temporal layers into its decoder and fine-tune these layers for higher temporal consistency. We also propose a set of strategies for composing video-text data that involve diverse content from multiple existing datasets including video datasets for action recognition and image-text datasets. Extensive experiments show that our method achieves good results in both quantitative and qualitative evaluations. Our project page is available https://anonymous0x233.github.io/ReuseAndDiffuse/{here}.

In-context learning for vision data has been underexplored compared with that in natural language. Previous works studied image in-context learning, urging models to generate a single image guided by demonstrations. In this paper, we propose and study video in-context learning, where the model starts from an existing video clip and generates diverse potential future sequences, each semantically guided by the prompted video demonstrations. To achieve this, we provide a clear definition of the task, and train an autoregressive Transformer on video datasets. We thoroughly analyze the effect of different datasets and represent frames as discrete tokens, and then model them by next token predictions. We design various evaluation metrics, including both objective and subjective measures, to demonstrate the visual quality and semantic accuracy of generation results. Our model follows the scaling law and generates high-quality video clips that accurately align with the semantic guidance provided by in-context examples.

Masked autoencoding has shown excellent performance on self-supervised video representation learning. Temporal redundancy has led to a high masking ratio and customized masking strategy in VideoMAE. In this paper, we aim to further improve the performance of video masked autoencoding by introducing a motion guided masking strategy. Our key insight is that motion is a general and unique prior in video, which should be taken into account during masked pre-training. Our motion guided masking explicitly incorporates motion information to build temporal consistent masking volume. Based on this masking volume, we can track the unmasked tokens in time and sample a set of temporal consistent cubes from videos. These temporal aligned unmasked tokens will further relieve the information leakage issue in time and encourage the MGMAE to learn more useful structure information. We implement our MGMAE with an online efficient optical flow estimator and backward masking map warping strategy. We perform experiments on the datasets of Something-Something V2 and Kinetics-400, demonstrating the superior performance of our MGMAE to the original VideoMAE. In addition, we provide the visualization analysis to illustrate that our MGMAE can sample temporal consistent cubes in a motion-adaptive manner for more effective video pre-training.

In this paper, we present StrucTexTv2, an effective document image pre-training framework, by performing masked visual-textual prediction. It consists of two self-supervised pre-training tasks: masked image modeling and masked language modeling, based on text region-level image masking. The proposed method randomly masks some image regions according to the bounding box coordinates of text words. The objectives of our pre-training tasks are reconstructing the pixels of masked image regions and the corresponding masked tokens simultaneously. Hence the pre-trained encoder can capture more textual semantics in comparison to the masked image modeling that usually predicts the masked image patches. Compared to the masked multi-modal modeling methods for document image understanding that rely on both the image and text modalities, StrucTexTv2 models image-only input and potentially deals with more application scenarios free from OCR pre-processing. Extensive experiments on mainstream benchmarks of document image understanding demonstrate the effectiveness of StrucTexTv2. It achieves competitive or even new state-of-the-art performance in various downstream tasks such as image classification, layout analysis, table structure recognition, document OCR, and information extraction under the end-to-end scenario.

Multimodal learning seeks to combine data from multiple input sources to enhance the performance of different downstream tasks. In real-world scenarios, performance can degrade substantially if some input modalities are missing. Existing methods that can handle missing modalities involve custom training or adaptation steps for each input modality combination. These approaches are either tied to specific modalities or become computationally expensive as the number of input modalities increases. In this paper, we propose Masked Modality Projection (MMP), a method designed to train a single model that is robust to any missing modality scenario. We achieve this by randomly masking a subset of modalities during training and learning to project available input modalities to estimate the tokens for the masked modalities. This approach enables the model to effectively learn to leverage the information from the available modalities to compensate for the missing ones, enhancing missing modality robustness. We conduct a series of experiments with various baseline models and datasets to assess the effectiveness of this strategy. Experiments demonstrate that our approach improves robustness to different missing modality scenarios, outperforming existing methods designed for missing modalities or specific modality combinations.

Self-supervised methods have shown remarkable progress in learning high-level semantics and low-level temporal correspondence. Building on these results, we take one step further and explore the possibility of integrating these two features to enhance object-centric representations. Our preliminary experiments indicate that query slot attention can extract different semantic components from the RGB feature map, while random sampling based slot attention can exploit temporal correspondence cues between frames to assist instance identification. Motivated by this, we propose a novel semantic-aware masked slot attention on top of the fused semantic features and correspondence maps. It comprises two slot attention stages with a set of shared learnable Gaussian distributions. In the first stage, we use the mean vectors as slot initialization to decompose potential semantics and generate semantic segmentation masks through iterative attention. In the second stage, for each semantics, we randomly sample slots from the corresponding Gaussian distribution and perform masked feature aggregation within the semantic area to exploit temporal correspondence patterns for instance identification. We adopt semantic- and instance-level temporal consistency as self-supervision to encourage temporally coherent object-centric representations. Our model effectively identifies multiple object instances with semantic structure, reaching promising results on unsupervised video object discovery. Furthermore, we achieve state-of-the-art performance on dense label propagation tasks, demonstrating the potential for object-centric analysis. The code is released at https://github.com/shvdiwnkozbw/SMTC.

Long-context autoregressive modeling has significantly advanced language generation, but video generation still struggles to fully utilize extended temporal contexts. To investigate long-context video modeling, we introduce Frame AutoRegressive (FAR), a strong baseline for video autoregressive modeling. Just as language models learn causal dependencies between tokens (i.e., Token AR), FAR models temporal causal dependencies between continuous frames, achieving better convergence than Token AR and video diffusion transformers. Building on FAR, we observe that long-context vision modeling faces challenges due to visual redundancy. Existing RoPE lacks effective temporal decay for remote context and fails to extrapolate well to long video sequences. Additionally, training on long videos is computationally expensive, as vision tokens grow much faster than language tokens. To tackle these issues, we propose balancing locality and long-range dependency. We introduce FlexRoPE, an test-time technique that adds flexible temporal decay to RoPE, enabling extrapolation to 16x longer vision contexts. Furthermore, we propose long short-term context modeling, where a high-resolution short-term context window ensures fine-grained temporal consistency, while an unlimited long-term context window encodes long-range information using fewer tokens. With this approach, we can train on long video sequences with a manageable token context length. We demonstrate that FAR achieves state-of-the-art performance in both short- and long-video generation, providing a simple yet effective baseline for video autoregressive modeling.

We present a neural network structure, FramePack, to train next-frame (or next-frame-section) prediction models for video generation. The FramePack compresses input frames to make the transformer context length a fixed number regardless of the video length. As a result, we are able to process a large number of frames using video diffusion with computation bottleneck similar to image diffusion. This also makes the training video batch sizes significantly higher (batch sizes become comparable to image diffusion training). We also propose an anti-drifting sampling method that generates frames in inverted temporal order with early-established endpoints to avoid exposure bias (error accumulation over iterations). Finally, we show that existing video diffusion models can be finetuned with FramePack, and their visual quality may be improved because the next-frame prediction supports more balanced diffusion schedulers with less extreme flow shift timesteps.

Scaling the size of a model enhances its capabilities but significantly increases computation complexity. Mixture-of-Experts models (MoE) address the issue by allowing model size to scale up without substantially increasing training or inference costs. Despite their promising results, MoE models encounter several challenges. Primarily, for dynamic routing methods, the dispersion of training tokens across multiple experts can lead to underfitting, particularly for infrequent tokens. Additionally, while fixed routing methods can mitigate that issue, they compromise on the diversity of representations. In this paper, we propose MaskMoE, a method designed to enhance token-level learning by employing a routing masking technique within the Mixture-of-Experts model. MaskMoE is capable of maintaining representation diversity while achieving more comprehensive training. Experimental results demonstrate that our method outperforms previous dominant Mixture-of-Experts models in terms of both perplexity (PPL) and downstream task performance.

Can we leverage the audiovisual information already present in video to improve self-supervised representation learning? To answer this question, we study various pretraining architectures and objectives within the masked autoencoding framework, motivated by the success of similar methods in natural language and image understanding. We show that we can achieve significant improvements on audiovisual downstream classification tasks, surpassing the state-of-the-art on VGGSound and AudioSet. Furthermore, we can leverage our audiovisual pretraining scheme for multiple unimodal downstream tasks using a single audiovisual pretrained model. We additionally demonstrate the transferability of our representations, achieving state-of-the-art audiovisual results on Epic Kitchens without pretraining specifically for this dataset.

Current vision models typically maintain a fixed correspondence between their representation structure and image space. Each layer comprises a set of tokens arranged "on-the-grid," which biases patches or tokens to encode information at a specific spatio(-temporal) location. In this work we present Moving Off-the-Grid (MooG), a self-supervised video representation model that offers an alternative approach, allowing tokens to move "off-the-grid" to better enable them to represent scene elements consistently, even as they move across the image plane through time. By using a combination of cross-attention and positional embeddings we disentangle the representation structure and image structure. We find that a simple self-supervised objective--next frame prediction--trained on video data, results in a set of latent tokens which bind to specific scene structures and track them as they move. We demonstrate the usefulness of MooG's learned representation both qualitatively and quantitatively by training readouts on top of the learned representation on a variety of downstream tasks. We show that MooG can provide a strong foundation for different vision tasks when compared to "on-the-grid" baselines.

Multimodal large language models (MLLMs) have revolutionized cross-modal understanding but continue to struggle with hallucinations - fabricated content contradicting visual inputs. Existing hallucination mitigation methods either incur prohibitive computational costs or introduce distribution mismatches between training data and model outputs. We identify a critical insight: hallucinations predominantly emerge at the early stages of text generation and propagate through subsequent outputs. To address this, we propose **SENTINEL** (**S**entence-level **E**arly i**N**tervention **T**hrough **IN**-domain pr**E**ference **L**earning), a framework that eliminates dependency on human annotations. Specifically, we first bootstrap high-quality in-domain preference pairs by iteratively sampling model outputs, validating object existence through cross-checking with two open-vocabulary detectors, and classifying sentences into hallucinated/non-hallucinated categories. Subsequently, we use context-coherent positive samples and hallucinated negative samples to build context-aware preference data iteratively. Finally, we train models using a context-aware preference loss (C-DPO) that emphasizes discriminative learning at the sentence level where hallucinations initially manifest. Experimental results show that SENTINEL can reduce hallucinations by over 90\% compared to the original model and outperforms the previous state-of-the-art method on both hallucination benchmarks and general capabilities benchmarks, demonstrating its superiority and generalization ability. The models, datasets, and code are available at https://github.com/pspdada/SENTINEL.

We present a meta-learning framework for learning new visual concepts quickly, from just one or a few examples, guided by multiple naturally occurring data streams: simultaneously looking at images, reading sentences that describe the objects in the scene, and interpreting supplemental sentences that relate the novel concept with other concepts. The learned concepts support downstream applications, such as answering questions by reasoning about unseen images. Our model, namely FALCON, represents individual visual concepts, such as colors and shapes, as axis-aligned boxes in a high-dimensional space (the "box embedding space"). Given an input image and its paired sentence, our model first resolves the referential expression in the sentence and associates the novel concept with particular objects in the scene. Next, our model interprets supplemental sentences to relate the novel concept with other known concepts, such as "X has property Y" or "X is a kind of Y". Finally, it infers an optimal box embedding for the novel concept that jointly 1) maximizes the likelihood of the observed instances in the image, and 2) satisfies the relationships between the novel concepts and the known ones. We demonstrate the effectiveness of our model on both synthetic and real-world datasets.

Masked AutoEncoders (MAE) have emerged as a robust self-supervised framework, offering remarkable performance across a wide range of downstream tasks. To increase the difficulty of the pretext task and learn richer visual representations, existing works have focused on replacing standard random masking with more sophisticated strategies, such as adversarial-guided and teacher-guided masking. However, these strategies depend on the input data thus commonly increasing the model complexity and requiring additional calculations to generate the mask patterns. This raises the question: Can we enhance MAE performance beyond random masking without relying on input data or incurring additional computational costs? In this work, we introduce a simple yet effective data-independent method, termed ColorMAE, which generates different binary mask patterns by filtering random noise. Drawing inspiration from color noise in image processing, we explore four types of filters to yield mask patterns with different spatial and semantic priors. ColorMAE requires no additional learnable parameters or computational overhead in the network, yet it significantly enhances the learned representations. We provide a comprehensive empirical evaluation, demonstrating our strategy's superiority in downstream tasks compared to random masking. Notably, we report an improvement of 2.72 in mIoU in semantic segmentation tasks relative to baseline MAE implementations.

Generalized few-shot semantic segmentation was introduced to move beyond only evaluating few-shot segmentation models on novel classes to include testing their ability to remember base classes. While the current state-of-the-art approach is based on meta-learning, it performs poorly and saturates in learning after observing only a few shots. We propose the first fine-tuning solution, and demonstrate that it addresses the saturation problem while achieving state-of-the-art results on two datasets, PASCAL-5i and COCO-20i. We also show that it outperforms existing methods, whether fine-tuning multiple final layers or only the final layer. Finally, we present a triplet loss regularization that shows how to redistribute the balance of performance between novel and base categories so that there is a smaller gap between them.

We use multilayer Long Short Term Memory (LSTM) networks to learn representations of video sequences. Our model uses an encoder LSTM to map an input sequence into a fixed length representation. This representation is decoded using single or multiple decoder LSTMs to perform different tasks, such as reconstructing the input sequence, or predicting the future sequence. We experiment with two kinds of input sequences - patches of image pixels and high-level representations ("percepts") of video frames extracted using a pretrained convolutional net. We explore different design choices such as whether the decoder LSTMs should condition on the generated output. We analyze the outputs of the model qualitatively to see how well the model can extrapolate the learned video representation into the future and into the past. We try to visualize and interpret the learned features. We stress test the model by running it on longer time scales and on out-of-domain data. We further evaluate the representations by finetuning them for a supervised learning problem - human action recognition on the UCF-101 and HMDB-51 datasets. We show that the representations help improve classification accuracy, especially when there are only a few training examples. Even models pretrained on unrelated datasets (300 hours of YouTube videos) can help action recognition performance.

We explore a new task for audio-visual-language modeling called fine-grained audible video description (FAVD). It aims to provide detailed textual descriptions for the given audible videos, including the appearance and spatial locations of each object, the actions of moving objects, and the sounds in videos. Existing visual-language modeling tasks often concentrate on visual cues in videos while undervaluing the language and audio modalities. On the other hand, FAVD requires not only audio-visual-language modeling skills but also paragraph-level language generation abilities. We construct the first fine-grained audible video description benchmark (FAVDBench) to facilitate this research. For each video clip, we first provide a one-sentence summary of the video, ie, the caption, followed by 4-6 sentences describing the visual details and 1-2 audio-related descriptions at the end. The descriptions are provided in both English and Chinese. We create two new metrics for this task: an EntityScore to gauge the completeness of entities in the visual descriptions, and an AudioScore to assess the audio descriptions. As a preliminary approach to this task, we propose an audio-visual-language transformer that extends existing video captioning model with an additional audio branch. We combine the masked language modeling and auto-regressive language modeling losses to optimize our model so that it can produce paragraph-level descriptions. We illustrate the efficiency of our model in audio-visual-language modeling by evaluating it against the proposed benchmark using both conventional captioning metrics and our proposed metrics. We further put our benchmark to the test in video generation models, demonstrating that employing fine-grained video descriptions can create more intricate videos than using captions.

Unifying visual understanding and generation within a single multimodal framework remains a significant challenge, as the two inherently heterogeneous tasks require representations at different levels of granularity. Current approaches that utilize vector quantization (VQ) or variational autoencoders (VAE) for unified visual representation prioritize intrinsic imagery features over semantics, compromising understanding performance. In this work, we take inspiration from masked image modelling (MIM) that learns rich semantics via a mask-and-reconstruct pre-training and its successful extension to masked autoregressive (MAR) image generation. A preliminary study on the MAR encoder's representation reveals exceptional linear probing accuracy and precise feature response to visual concepts, which indicates MAR's potential for visual understanding tasks beyond its original generation role. Based on these insights, we present Harmon, a unified autoregressive framework that harmonizes understanding and generation tasks with a shared MAR encoder. Through a three-stage training procedure that progressively optimizes understanding and generation capabilities, Harmon achieves state-of-the-art image generation results on the GenEval, MJHQ30K and WISE benchmarks while matching the performance of methods with dedicated semantic encoders (e.g., Janus) on image understanding benchmarks. Our code and models will be available at https://github.com/wusize/Harmon.

The increasing demand to process long and high-resolution videos significantly burdens Large Vision-Language Models (LVLMs) due to the enormous number of visual tokens. Existing token reduction methods primarily focus on importance-based token pruning, which overlooks the redundancy caused by frame resemblance and repetitive visual elements. In this paper, we analyze the high vision token similarities in LVLMs. We reveal that token similarity distribution condenses as layers deepen while maintaining ranking consistency. Leveraging the unique properties of similarity over importance, we introduce FrameFusion, a novel approach that combines similarity-based merging with importance-based pruning for better token reduction in LVLMs. FrameFusion identifies and merges similar tokens before pruning, opening up a new perspective for token reduction. We evaluate FrameFusion on diverse LVLMs, including Llava-Video-{7B,32B,72B}, and MiniCPM-V-8B, on video understanding, question-answering, and retrieval benchmarks. Experiments show that FrameFusion reduces vision tokens by 70%, achieving 3.4-4.4x LLM speedups and 1.6-1.9x end-to-end speedups, with an average performance impact of less than 3%. Our code is available at https://github.com/thu-nics/FrameFusion.

Sequential video understanding, as an emerging video understanding task, has driven lots of researchers' attention because of its goal-oriented nature. This paper studies weakly supervised sequential video understanding where the accurate time-stamp level text-video alignment is not provided. We solve this task by borrowing ideas from CLIP. Specifically, we use a transformer to aggregate frame-level features for video representation and use a pre-trained text encoder to encode the texts corresponding to each action and the whole video, respectively. To model the correspondence between text and video, we propose a multiple granularity loss, where the video-paragraph contrastive loss enforces matching between the whole video and the complete script, and a fine-grained frame-sentence contrastive loss enforces the matching between each action and its description. As the frame-sentence correspondence is not available, we propose to use the fact that video actions happen sequentially in the temporal domain to generate pseudo frame-sentence correspondence and supervise the network training with the pseudo labels. Extensive experiments on video sequence verification and text-to-video matching show that our method outperforms baselines by a large margin, which validates the effectiveness of our proposed approach. Code is available at https://github.com/svip-lab/WeakSVR

Despite recent advances in Large Video Language Models (LVLMs), they still struggle with fine-grained temporal understanding, hallucinate, and often make simple mistakes on even simple video question-answering tasks, all of which pose significant challenges to their safe and reliable deployment in real-world applications. To address these limitations, we propose a self-alignment framework that enables LVLMs to learn from their own errors. Our proposed framework first obtains a training set of preferred and non-preferred response pairs, where non-preferred responses are generated by incorporating common error patterns that often occur due to inadequate spatio-temporal understanding, spurious correlations between co-occurring concepts, and over-reliance on linguistic cues while neglecting the vision modality, among others. To facilitate self-alignment of LVLMs with the constructed preferred and non-preferred response pairs, we introduce Refined Regularized Preference Optimization (RRPO), a novel preference optimization method that utilizes sub-sequence-level refined rewards and token-wise KL regularization to address the limitations of Direct Preference Optimization (DPO). We demonstrate that RRPO achieves more precise alignment and more stable training compared to DPO. Our experiments and analysis validate the effectiveness of our approach across diverse video tasks, including video hallucination, short- and long-video understanding, and fine-grained temporal reasoning.

The rapid progress of auto-regressive vision-language models (VLMs) has inspired growing interest in vision-language-action models (VLA) for robotic manipulation. Recently, masked diffusion models, a paradigm distinct from autoregressive models, have begun to demonstrate competitive performance in text generation and multimodal applications, leading to the development of a series of diffusion-based VLMs (d-VLMs). However, leveraging such models for robot policy learning remains largely unexplored. In this work, we present LLaDA-VLA, the first Vision-Language-Diffusion-Action model built upon pretrained d-VLMs for robotic manipulation. To effectively adapt d-VLMs to robotic domain, we introduce two key designs: (1) a localized special-token classification strategy that replaces full-vocabulary classification with special action token classification, reducing adaptation difficulty; (2) a hierarchical action-structured decoding strategy that decodes action sequences hierarchically considering the dependencies within and across actions. Extensive experiments demonstrate that LLaDA-VLA significantly outperforms state-of-the-art VLAs on both simulation and real-world robots.

Video Anomaly Detection (VAD) methods based on reconstruction or prediction face two critical challenges: (1) strong generalization capability often results in accurate reconstruction or prediction of abnormal events, making it difficult to distinguish normal from abnormal patterns; (2) reliance only on low-level appearance and motion cues limits their ability to identify high-level semantic in abnormal events from complex scenes. To address these limitations, we propose a novel VAD framework with two key innovations. First, to suppress excessive generalization, we introduce the Sparse Feature Filtering Module (SFFM) that employs bottleneck filters to dynamically and adaptively remove abnormal information from features. Unlike traditional memory modules, it does not need to memorize the normal prototypes across the training dataset. Further, we design the Mixture of Experts (MoE) architecture for SFFM. Each expert is responsible for extracting specialized principal features during running time, and different experts are selectively activated to ensure the diversity of the learned principal features. Second, to overcome the neglect of semantics in existing methods, we integrate a Vision-Language Model (VLM) to generate textual descriptions for video clips, enabling comprehensive joint modeling of semantic, appearance, and motion cues. Additionally, we enforce modality consistency through semantic similarity constraints and motion frame-difference contrastive loss. Extensive experiments on multiple public datasets validate the effectiveness of our multimodal joint modeling framework and sparse feature filtering paradigm. Project page at https://qzfm.github.io/sfn_vad_project_page/.

We introduce a hierarchical filtering procedure for generative models of sequences on trees, enabling control over the range of positional correlations in the data. Leveraging this controlled setting, we provide evidence that vanilla encoder-only transformer architectures can implement the optimal Belief Propagation algorithm on both root classification and masked language modeling tasks. Correlations at larger distances corresponding to increasing layers of the hierarchy are sequentially included as the network is trained. We analyze how the transformer layers succeed by focusing on attention maps from models trained with varying degrees of filtering. These attention maps show clear evidence for iterative hierarchical reconstruction of correlations, and we can relate these observations to a plausible implementation of the exact inference algorithm for the network sizes considered.

Unified multimodal models (UMMs) have emerged as a powerful paradigm in foundational computer vision research, demonstrating significant potential in both image understanding and generation. However, existing research in the face domain primarily focuses on coarse facial attribute understanding, with limited capacity to handle fine-grained facial attributes and without addressing generation capabilities. To overcome these limitations, we propose UniF^2ace, the first UMM tailored specifically for fine-grained face understanding and generation. In general, we train UniF^2ace on a self-constructed, specialized dataset utilizing two mutually beneficial diffusion techniques and a two-level mixture-of-experts architecture. Specifically, we first build a large-scale facial dataset, UniF^2ace-130K, which contains 130K image-text pairs with one million question-answering pairs that span a wide range of facial attributes. Second, we establish a theoretical connection between discrete diffusion score matching and masked generative models, optimizing both evidence lower bounds simultaneously, which significantly improves the model's ability to synthesize facial details. Finally, we introduce both token-level and sequence-level mixture-of-experts, enabling efficient fine-grained representation learning for both understanding and generation tasks. Extensive experiments on UniF^2ace-130K demonstrate that UniF^2ace outperforms existing UMMs and generative models, achieving superior performance across both understanding and generation tasks.

How to effectively explore spatial-temporal features is important for video colorization. Instead of stacking multiple frames along the temporal dimension or recurrently propagating estimated features that will accumulate errors or cannot explore information from far-apart frames, we develop a memory-based feature propagation module that can establish reliable connections with features from far-apart frames and alleviate the influence of inaccurately estimated features. To extract better features from each frame for the above-mentioned feature propagation, we explore the features from large-pretrained visual models to guide the feature estimation of each frame so that the estimated features can model complex scenarios. In addition, we note that adjacent frames usually contain similar contents. To explore this property for better spatial and temporal feature utilization, we develop a local attention module to aggregate the features from adjacent frames in a spatial-temporal neighborhood. We formulate our memory-based feature propagation module, large-pretrained visual model guided feature estimation module, and local attention module into an end-to-end trainable network (named ColorMNet) and show that it performs favorably against state-of-the-art methods on both the benchmark datasets and real-world scenarios. The source code and pre-trained models will be available at https://github.com/yyang181/colormnet.

In this paper we present a new computer vision task, named video instance segmentation. The goal of this new task is simultaneous detection, segmentation and tracking of instances in videos. In words, it is the first time that the image instance segmentation problem is extended to the video domain. To facilitate research on this new task, we propose a large-scale benchmark called YouTube-VIS, which consists of 2883 high-resolution YouTube videos, a 40-category label set and 131k high-quality instance masks. In addition, we propose a novel algorithm called MaskTrack R-CNN for this task. Our new method introduces a new tracking branch to Mask R-CNN to jointly perform the detection, segmentation and tracking tasks simultaneously. Finally, we evaluate the proposed method and several strong baselines on our new dataset. Experimental results clearly demonstrate the advantages of the proposed algorithm and reveal insight for future improvement. We believe the video instance segmentation task will motivate the community along the line of research for video understanding.

Diffusion models (DMs) have become the new trend of generative models and have demonstrated a powerful ability of conditional synthesis. Among those, text-to-image diffusion models pre-trained on large-scale image-text pairs are highly controllable by customizable prompts. Unlike the unconditional generative models that focus on low-level attributes and details, text-to-image diffusion models contain more high-level knowledge thanks to the vision-language pre-training. In this paper, we propose VPD (Visual Perception with a pre-trained Diffusion model), a new framework that exploits the semantic information of a pre-trained text-to-image diffusion model in visual perception tasks. Instead of using the pre-trained denoising autoencoder in a diffusion-based pipeline, we simply use it as a backbone and aim to study how to take full advantage of the learned knowledge. Specifically, we prompt the denoising decoder with proper textual inputs and refine the text features with an adapter, leading to a better alignment to the pre-trained stage and making the visual contents interact with the text prompts. We also propose to utilize the cross-attention maps between the visual features and the text features to provide explicit guidance. Compared with other pre-training methods, we show that vision-language pre-trained diffusion models can be faster adapted to downstream visual perception tasks using the proposed VPD. Extensive experiments on semantic segmentation, referring image segmentation and depth estimation demonstrates the effectiveness of our method. Notably, VPD attains 0.254 RMSE on NYUv2 depth estimation and 73.3% oIoU on RefCOCO-val referring image segmentation, establishing new records on these two benchmarks. Code is available at https://github.com/wl-zhao/VPD

Masked Autoencoder (MAE) is a self-supervised approach for representation learning, widely applicable to a variety of downstream tasks in computer vision. In spite of its success, it is still not fully uncovered what and how MAE exactly learns. In this paper, with an in-depth analysis, we discover that MAE intrinsically learns pattern-based patch-level clustering from surprisingly early stages of pretraining. Upon this understanding, we propose self-guided masked autoencoder, which internally generates informed mask by utilizing its progress in patch clustering, substituting the naive random masking of the vanilla MAE. Our approach significantly boosts its learning process without relying on any external models or supplementary information, keeping the benefit of self-supervised nature of MAE intact. Comprehensive experiments on various downstream tasks verify the effectiveness of the proposed method.

Videos are highly redundant data source and it is often enough to identify a few key moments to solve any given task. In this paper, we present a text-conditioned video resampler (TCR) module that uses a pre-trained and frozen visual encoder and large language model (LLM) to process long video sequences for a task. TCR localises relevant visual features from the video given a text condition and provides them to a LLM to generate a text response. Due to its lightweight design and use of cross-attention, TCR can process more than 100 frames at a time allowing the model to use much longer chunks of video than earlier works. We make the following contributions: (i) we design a transformer-based sampling architecture that can process long videos conditioned on a task, together with a training method that enables it to bridge pre-trained visual and language models; (ii) we empirically validate its efficacy on a wide variety of evaluation tasks, and set a new state-of-the-art on NextQA, EgoSchema, and the EGO4D-LTA challenge; and (iii) we determine tasks which require longer video contexts and that can thus be used effectively for further evaluation of long-range video models.

Scale is the primary factor for building a powerful foundation model that could well generalize to a variety of downstream tasks. However, it is still challenging to train video foundation models with billions of parameters. This paper shows that video masked autoencoder (VideoMAE) is a scalable and general self-supervised pre-trainer for building video foundation models. We scale the VideoMAE in both model and data with a core design. Specifically, we present a dual masking strategy for efficient pre-training, with an encoder operating on a subset of video tokens and a decoder processing another subset of video tokens. Although VideoMAE is very efficient due to high masking ratio in encoder, masking decoder can still further reduce the overall computational cost. This enables the efficient pre-training of billion-level models in video. We also use a progressive training paradigm that involves an initial pre-training on a diverse multi-sourced unlabeled dataset, followed by a post-pre-training on a mixed labeled dataset. Finally, we successfully train a video ViT model with a billion parameters, which achieves a new state-of-the-art performance on the datasets of Kinetics (90.0% on K400 and 89.9% on K600) and Something-Something (68.7% on V1 and 77.0% on V2). In addition, we extensively verify the pre-trained video ViT models on a variety of downstream tasks, demonstrating its effectiveness as a general video representation learner. The code and model is available at https://github.com/OpenGVLab/VideoMAEv2.

We present a unified, promptable model capable of simultaneously segmenting, recognizing, and captioning anything. Unlike SAM, we aim to build a versatile region representation in the wild via visual prompting. To achieve this, we train a generalizable model with massive segmentation masks, e.g., SA-1B masks, and semantic priors from a pre-trained CLIP model with 5 billion parameters. Specifically, we construct a promptable image decoder by adding a semantic token to each mask token. The semantic token is responsible for learning the semantic priors in a predefined concept space. Through joint optimization of segmentation on mask tokens and concept prediction on semantic tokens, our model exhibits strong regional recognition and localization capabilities. For example, an additional 38M-parameter causal text decoder trained from scratch sets a new record with a CIDEr score of 150.7 on the Visual Genome region captioning task. We believe this model can be a versatile region-level image tokenizer, capable of encoding general-purpose region context for a broad range of perception tasks. Code and models are available at https://github.com/baaivision/tokenize-anything.

The self-supervised objective of masking-and-predicting has led to promising performance gains on a variety of downstream tasks. However, while most approaches randomly mask tokens, there is strong intuition that deciding what to mask can substantially improve learning outcomes. We investigate this in continued pretraining setting in which pretrained models continue to pretrain on domain-specific data before performing some downstream task. We introduce Difference-Masking, a masking strategy that automatically chooses what to mask during continued pretraining by considering what makes a task domain different from the pretraining domain. Empirically, we find that Difference-Masking outperforms baselines on continued pretraining settings across four diverse language-only and multimodal video tasks.

We consider the question: what can be learnt by looking at and listening to a large number of unlabelled videos? There is a valuable, but so far untapped, source of information contained in the video itself -- the correspondence between the visual and the audio streams, and we introduce a novel "Audio-Visual Correspondence" learning task that makes use of this. Training visual and audio networks from scratch, without any additional supervision other than the raw unconstrained videos themselves, is shown to successfully solve this task, and, more interestingly, result in good visual and audio representations. These features set the new state-of-the-art on two sound classification benchmarks, and perform on par with the state-of-the-art self-supervised approaches on ImageNet classification. We also demonstrate that the network is able to localize objects in both modalities, as well as perform fine-grained recognition tasks.

Contrastive language-image pretraining has shown great success in learning visual-textual joint representation from web-scale data, demonstrating remarkable "zero-shot" generalization ability for various image tasks. However, how to effectively expand such new language-image pretraining methods to video domains is still an open problem. In this work, we present a simple yet effective approach that adapts the pretrained language-image models to video recognition directly, instead of pretraining a new model from scratch. More concretely, to capture the long-range dependencies of frames along the temporal dimension, we propose a cross-frame attention mechanism that explicitly exchanges information across frames. Such module is lightweight and can be plugged into pretrained language-image models seamlessly. Moreover, we propose a video-specific prompting scheme, which leverages video content information for generating discriminative textual prompts. Extensive experiments demonstrate that our approach is effective and can be generalized to different video recognition scenarios. In particular, under fully-supervised settings, our approach achieves a top-1 accuracy of 87.1% on Kinectics-400, while using 12 times fewer FLOPs compared with Swin-L and ViViT-H. In zero-shot experiments, our approach surpasses the current state-of-the-art methods by +7.6% and +14.9% in terms of top-1 accuracy under two popular protocols. In few-shot scenarios, our approach outperforms previous best methods by +32.1% and +23.1% when the labeled data is extremely limited. Code and models are available at https://aka.ms/X-CLIP

We introduce a novel sequential modeling approach which enables learning a Large Vision Model (LVM) without making use of any linguistic data. To do this, we define a common format, "visual sentences", in which we can represent raw images and videos as well as annotated data sources such as semantic segmentations and depth reconstructions without needing any meta-knowledge beyond the pixels. Once this wide variety of visual data (comprising 420 billion tokens) is represented as sequences, the model can be trained to minimize a cross-entropy loss for next token prediction. By training across various scales of model architecture and data diversity, we provide empirical evidence that our models scale effectively. Many different vision tasks can be solved by designing suitable visual prompts at test time.

We propose and investigate an identity sensitive joint embedding of face and voice. Such an embedding enables cross-modal retrieval from voice to face and from face to voice. We make the following four contributions: first, we show that the embedding can be learnt from videos of talking faces, without requiring any identity labels, using a form of cross-modal self-supervision; second, we develop a curriculum learning schedule for hard negative mining targeted to this task, that is essential for learning to proceed successfully; third, we demonstrate and evaluate cross-modal retrieval for identities unseen and unheard during training over a number of scenarios and establish a benchmark for this novel task; finally, we show an application of using the joint embedding for automatically retrieving and labelling characters in TV dramas.

We present a simplified, task-agnostic multi-modal pre-training approach that can accept either video or text input, or both for a variety of end tasks. Existing pre-training are task-specific by adopting either a single cross-modal encoder that requires both modalities, limiting their use for retrieval-style end tasks or more complex multitask learning with two unimodal encoders, limiting early cross-modal fusion. We instead introduce new pretraining masking schemes that better mix across modalities (e.g. by forcing masks for text to predict the closest video embeddings) while also maintaining separability (e.g. unimodal predictions are sometimes required, without using all the input). Experimental results show strong performance across a wider range of tasks than any previous methods, often outperforming task-specific pre-training. Code is made available at https://github.com/pytorch/fairseq/tree/main/examples/MMPT.

Recent approaches have shown that large-scale vision-language models such as CLIP can improve semantic segmentation performance. These methods typically aim for pixel-level vision-language alignment, but often rely on low resolution image features from CLIP, resulting in class ambiguities along boundaries. Moreover, the global scene representations in CLIP text embeddings do not directly correlate with the local and detailed pixel-level features, making meaningful alignment more difficult. To address these limitations, we introduce MTA-CLIP, a novel framework employing mask-level vision-language alignment. Specifically, we first propose Mask-Text Decoder that enhances the mask representations using rich textual data with the CLIP language model. Subsequently, it aligns mask representations with text embeddings using Mask-to-Text Contrastive Learning. Furthermore, we introduce MaskText Prompt Learning, utilizing multiple context-specific prompts for text embeddings to capture diverse class representations across masks. Overall, MTA-CLIP achieves state-of-the-art, surpassing prior works by an average of 2.8% and 1.3% on on standard benchmark datasets, ADE20k and Cityscapes, respectively.

In this paper, we present IMAGEdit, a training-free framework for any number of video subject editing that manipulates the appearances of multiple designated subjects while preserving non-target regions, without finetuning or retraining. We achieve this by providing robust multimodal conditioning and precise mask sequences through a prompt-guided multimodal alignment module and a prior-based mask retargeting module. We first leverage large models' understanding and generation capabilities to produce multimodal information and mask motion sequences for multiple subjects across various types. Then, the obtained prior mask sequences are fed into a pretrained mask-driven video generation model to synthesize the edited video. With strong generalization capability, IMAGEdit remedies insufficient prompt-side multimodal conditioning and overcomes mask boundary entanglement in videos with any number of subjects, thereby significantly expanding the applicability of video editing. More importantly, IMAGEdit is compatible with any mask-driven video generation model, significantly improving overall performance. Extensive experiments on our newly constructed multi-subject benchmark MSVBench verify that IMAGEdit consistently surpasses state-of-the-art methods. Code, models, and datasets are publicly available at https://github.com/XWH-A/IMAGEdit.

The objective of this paper is an automatic Audio Description (AD) model that ingests movies and outputs AD in text form. Generating high-quality movie AD is challenging due to the dependency of the descriptions on context, and the limited amount of training data available. In this work, we leverage the power of pretrained foundation models, such as GPT and CLIP, and only train a mapping network that bridges the two models for visually-conditioned text generation. In order to obtain high-quality AD, we make the following four contributions: (i) we incorporate context from the movie clip, AD from previous clips, as well as the subtitles; (ii) we address the lack of training data by pretraining on large-scale datasets, where visual or contextual information is unavailable, e.g. text-only AD without movies or visual captioning datasets without context; (iii) we improve on the currently available AD datasets, by removing label noise in the MAD dataset, and adding character naming information; and (iv) we obtain strong results on the movie AD task compared with previous methods.

Dual encoder architectures like CLIP models map two types of inputs into a shared embedding space and predict similarities between them. Despite their success, it is, however, not understood how these models compare their two inputs. Common first-order feature-attribution methods can only provide limited insights into dual-encoders since their predictions depend on feature-interactions rather than on individual features. In this paper, we first derive a second-order method enabling the attribution of predictions by any differentiable dual encoder onto feature-interactions between its inputs. Second, we apply our method to CLIP models and show that they learn fine-grained correspondences between parts of captions and regions in images. They match objects across input modes also account for mismatches. This visual-linguistic grounding ability, however, varies heavily between object classes and exhibits pronounced out-of-domain effects. We can identify individual errors as well as systematic failure categories including object coverage, unusual scenes and correlated contexts.

Recently, Large Vision-Language Models (LVLMs) have demonstrated impressive capabilities in multi-modal context comprehension. However, they still suffer from hallucination problems referring to generating inconsistent outputs with the image content. To mitigate hallucinations, previous studies mainly focus on retraining LVLMs with custom datasets. Although effective, they inherently come with additional computational costs. In this paper, we propose a training-free framework, MVP, that aims to reduce hallucinations by making the most of the innate capabilities of the LVLMs via Multi-View Multi-Path Reasoning. Specifically, we first devise a multi-view information-seeking strategy to thoroughly perceive the comprehensive information in the image, which enriches the general global information captured by the original vision encoder in LVLMs. Furthermore, during the answer decoding, we observe that the occurrence of hallucinations has a strong correlation with the certainty of the answer tokens. Thus, we propose multi-path reasoning for each information view to quantify and aggregate the certainty scores for each potential answer among multiple decoding paths and finally decide the output answer. By fully grasping the information in the image and carefully considering the certainty of the potential answers when decoding, our MVP can effectively reduce hallucinations in LVLMs.The extensive experiments verify that our proposed MVP significantly mitigates the hallucination problem across four well-known LVLMs. The source code is available at: https://github.com/GasolSun36/MVP.

We introduce nested diffusion models, an efficient and powerful hierarchical generative framework that substantially enhances the generation quality of diffusion models, particularly for images of complex scenes. Our approach employs a series of diffusion models to progressively generate latent variables at different semantic levels. Each model in this series is conditioned on the output of the preceding higher-level models, culminating in image generation. Hierarchical latent variables guide the generation process along predefined semantic pathways, allowing our approach to capture intricate structural details while significantly improving image quality. To construct these latent variables, we leverage a pre-trained visual encoder, which learns strong semantic visual representations, and modulate its capacity via dimensionality reduction and noise injection. Across multiple datasets, our system demonstrates significant enhancements in image quality for both unconditional and class/text conditional generation. Moreover, our unconditional generation system substantially outperforms the baseline conditional system. These advancements incur minimal computational overhead as the more abstract levels of our hierarchy work with lower-dimensional representations.

We introduce a novel self-supervised learning approach to learn representations of videos that are responsive to changes in the motion dynamics. Our representations can be learned from data without human annotation and provide a substantial boost to the training of neural networks on small labeled data sets for tasks such as action recognition, which require to accurately distinguish the motion of objects. We promote an accurate learning of motion without human annotation by training a neural network to discriminate a video sequence from its temporally transformed versions. To learn to distinguish non-trivial motions, the design of the transformations is based on two principles: 1) To define clusters of motions based on time warps of different magnitude; 2) To ensure that the discrimination is feasible only by observing and analyzing as many image frames as possible. Thus, we introduce the following transformations: forward-backward playback, random frame skipping, and uniform frame skipping. Our experiments show that networks trained with the proposed method yield representations with improved transfer performance for action recognition on UCF101 and HMDB51.

The rapid development of Large Multimodal Models (LMMs) has significantly advanced multimodal understanding by harnessing the language abilities of Large Language Models (LLMs) and integrating modality-specific encoders. However, LMMs are plagued by hallucinations that limit their reliability and adoption. While traditional methods to detect and mitigate these hallucinations often involve costly training or rely heavily on external models, recent approaches utilizing internal model features present a promising alternative. In this paper, we critically assess the limitations of the state-of-the-art training-free technique, the logit lens, in handling generalized visual hallucinations. We introduce a refined method that leverages contextual token embeddings from middle layers of LMMs. This approach significantly improves hallucination detection and grounding across diverse categories, including actions and OCR, while also excelling in tasks requiring contextual understanding, such as spatial relations and attribute comparison. Our novel grounding technique yields highly precise bounding boxes, facilitating a transition from Zero-Shot Object Segmentation to Grounded Visual Question Answering. Our contributions pave the way for more reliable and interpretable multimodal models.

Model pre-training is essential in human-centric perception. In this paper, we first introduce masked image modeling (MIM) as a pre-training approach for this task. Upon revisiting the MIM training strategy, we reveal that human structure priors offer significant potential. Motivated by this insight, we further incorporate an intuitive human structure prior - human parts - into pre-training. Specifically, we employ this prior to guide the mask sampling process. Image patches, corresponding to human part regions, have high priority to be masked out. This encourages the model to concentrate more on body structure information during pre-training, yielding substantial benefits across a range of human-centric perception tasks. To further capture human characteristics, we propose a structure-invariant alignment loss that enforces different masked views, guided by the human part prior, to be closely aligned for the same image. We term the entire method as HAP. HAP simply uses a plain ViT as the encoder yet establishes new state-of-the-art performance on 11 human-centric benchmarks, and on-par result on one dataset. For example, HAP achieves 78.1% mAP on MSMT17 for person re-identification, 86.54% mA on PA-100K for pedestrian attribute recognition, 78.2% AP on MS COCO for 2D pose estimation, and 56.0 PA-MPJPE on 3DPW for 3D pose and shape estimation.

With the development of video generation models has advanced significantly in recent years, we adopt large-scale image-to-video diffusion models for video frame interpolation. We present a conditional encoder designed to adapt an image-to-video model for large-motion frame interpolation. To enhance performance, we integrate a dual-branch feature extractor and propose a cross-frame attention mechanism that effectively captures both spatial and temporal information, enabling accurate interpolations of intermediate frames. Our approach demonstrates superior performance on the Fr\'echet Video Distance (FVD) metric when evaluated against other state-of-the-art approaches, particularly in handling large motion scenarios, highlighting advancements in generative-based methodologies.

Unifying diverse image generation tasks within a single framework remains a fundamental challenge in visual generation. While large language models (LLMs) achieve unification through task-agnostic data and generation, existing visual generation models fail to meet these principles. Current approaches either rely on per-task datasets and large-scale training or adapt pre-trained image models with task-specific modifications, limiting their generalizability. In this work, we explore video models as a foundation for unified image generation, leveraging their inherent ability to model temporal correlations. We introduce RealGeneral, a novel framework that reformulates image generation as a conditional frame prediction task, analogous to in-context learning in LLMs. To bridge the gap between video models and condition-image pairs, we propose (1) a Unified Conditional Embedding module for multi-modal alignment and (2) a Unified Stream DiT Block with decoupled adaptive LayerNorm and attention mask to mitigate cross-modal interference. RealGeneral demonstrates effectiveness in multiple important visual generation tasks, e.g., it achieves a 14.5% improvement in subject similarity for customized generation and a 10% enhancement in image quality for canny-to-image task. Project page: https://lyne1.github.io/RealGeneral/

We present Phenaki, a model capable of realistic video synthesis, given a sequence of textual prompts. Generating videos from text is particularly challenging due to the computational cost, limited quantities of high quality text-video data and variable length of videos. To address these issues, we introduce a new model for learning video representation which compresses the video to a small representation of discrete tokens. This tokenizer uses causal attention in time, which allows it to work with variable-length videos. To generate video tokens from text we are using a bidirectional masked transformer conditioned on pre-computed text tokens. The generated video tokens are subsequently de-tokenized to create the actual video. To address data issues, we demonstrate how joint training on a large corpus of image-text pairs as well as a smaller number of video-text examples can result in generalization beyond what is available in the video datasets. Compared to the previous video generation methods, Phenaki can generate arbitrary long videos conditioned on a sequence of prompts (i.e. time variable text or a story) in open domain. To the best of our knowledge, this is the first time a paper studies generating videos from time variable prompts. In addition, compared to the per-frame baselines, the proposed video encoder-decoder computes fewer tokens per video but results in better spatio-temporal consistency.

Recent advances in Large Multimodal Models (LMMs) lead to significant breakthroughs in both academia and industry. One question that arises is how we, as humans, can understand their internal neural representations. This paper takes an initial step towards addressing this question by presenting a versatile framework to identify and interpret the semantics within LMMs. Specifically, 1) we first apply a Sparse Autoencoder(SAE) to disentangle the representations into human understandable features. 2) We then present an automatic interpretation framework to interpreted the open-semantic features learned in SAE by the LMMs themselves. We employ this framework to analyze the LLaVA-NeXT-8B model using the LLaVA-OV-72B model, demonstrating that these features can effectively steer the model's behavior. Our results contribute to a deeper understanding of why LMMs excel in specific tasks, including EQ tests, and illuminate the nature of their mistakes along with potential strategies for their rectification. These findings offer new insights into the internal mechanisms of LMMs and suggest parallels with the cognitive processes of the human brain.

This study addresses the issue observed in Large Vision Language Models (LVLMs), where excessive attention on a few image tokens, referred to as blind tokens, leads to hallucinatory responses in tasks requiring fine-grained understanding of visual objects. We found that tokens receiving lower attention weights often hold essential information for identifying nuanced object details -- ranging from merely recognizing object existence to identifying their attributes (color, position, etc.) and understanding their relationships. To counteract the over-emphasis on blind tokens and to accurately respond to user queries, we introduce a technique called Attentional Vision Calibration (AVC). During the decoding phase, AVC identifies blind tokens by analyzing the image-related attention distribution. It then dynamically adjusts the logits for the next token prediction by contrasting the logits conditioned on the original visual tokens with those conditioned on the blind tokens. This effectively lowers the dependency on blind tokens and promotes a more balanced consideration of all tokens. We validate AVC on benchmarks such as POPE, MME, and AMBER, where it consistently outperforms existing decoding techniques in mitigating object hallucinations in LVLMs.

Long video question answering is a challenging task that involves recognizing short-term activities and reasoning about their fine-grained relationships. State-of-the-art video Large Language Models (vLLMs) hold promise as a viable solution due to their demonstrated emergent capabilities on new tasks. However, despite being trained on millions of short seconds-long videos, vLLMs are unable to understand minutes-long videos and accurately answer questions about them. To address this limitation, we propose a lightweight and self-supervised approach, Key frame-conditioned long video-LLM (Koala), that introduces learnable spatiotemporal queries to adapt pretrained vLLMs for generalizing to longer videos. Our approach introduces two new tokenizers that condition on visual tokens computed from sparse video key frames for understanding short and long video moments. We train our proposed approach on HowTo100M and demonstrate its effectiveness on zero-shot long video understanding benchmarks, where it outperforms state-of-the-art large models by 3 - 6% in absolute accuracy across all tasks. Surprisingly, we also empirically show that our approach not only helps a pretrained vLLM to understand long videos but also improves its accuracy on short-term action recognition.

In text-video retrieval, recent works have benefited from the powerful learning capabilities of pre-trained text-image foundation models (e.g., CLIP) by adapting them to the video domain. A critical problem for them is how to effectively capture the rich semantics inside the video using the image encoder of CLIP. To tackle this, state-of-the-art methods adopt complex cross-modal modeling techniques to fuse the text information into video frame representations, which, however, incurs severe efficiency issues in large-scale retrieval systems as the video representations must be recomputed online for every text query. In this paper, we discard this problematic cross-modal fusion process and aim to learn semantically-enhanced representations purely from the video, so that the video representations can be computed offline and reused for different texts. Concretely, we first introduce a spatial-temporal "Prompt Cube" into the CLIP image encoder and iteratively switch it within the encoder layers to efficiently incorporate the global video semantics into frame representations. We then propose to apply an auxiliary video captioning objective to train the frame representations, which facilitates the learning of detailed video semantics by providing fine-grained guidance in the semantic space. With a naive temporal fusion strategy (i.e., mean-pooling) on the enhanced frame representations, we obtain state-of-the-art performances on three benchmark datasets, i.e., MSR-VTT, MSVD, and LSMDC.

This paper presents that the masked-modeling principle driving the success of large foundational vision models can be effectively applied to audio by making predictions in a latent space. We introduce Audio-based Joint-Embedding Predictive Architecture (A-JEPA), a simple extension method for self-supervised learning from the audio spectrum. Following the design of I-JEPA, our A-JEPA encodes visible audio spectrogram patches with a curriculum masking strategy via context encoder, and predicts the representations of regions sampled at well-designed locations. The target representations of those regions are extracted by the exponential moving average of context encoder, i.e., target encoder, on the whole spectrogram. We find it beneficial to transfer random block masking into time-frequency aware masking in a curriculum manner, considering the complexity of highly correlated in local time and frequency in audio spectrograms. To enhance contextual semantic understanding and robustness, we fine-tune the encoder with a regularized masking on target datasets, instead of input dropping or zero. Empirically, when built with Vision Transformers structure, we find A-JEPA to be highly scalable and sets new state-of-the-art performance on multiple audio and speech classification tasks, outperforming other recent models that use externally supervised pre-training.

Video understanding represents the most challenging frontier in computer vision, requiring models to reason about complex spatiotemporal relationships, long-term dependencies, and multimodal evidence. The recent emergence of Video-Large Multimodal Models (Video-LMMs), which integrate visual encoders with powerful decoder-based language models, has demonstrated remarkable capabilities in video understanding tasks. However, the critical phase that transforms these models from basic perception systems into sophisticated reasoning engines, post-training, remains fragmented across the literature. This survey provides the first comprehensive examination of post-training methodologies for Video-LMMs, encompassing three fundamental pillars: supervised fine-tuning (SFT) with chain-of-thought, reinforcement learning (RL) from verifiable objectives, and test-time scaling (TTS) through enhanced inference computation. We present a structured taxonomy that clarifies the roles, interconnections, and video-specific adaptations of these techniques, addressing unique challenges such as temporal localization, spatiotemporal grounding, long video efficiency, and multimodal evidence integration. Through systematic analysis of representative methods, we synthesize key design principles, insights, and evaluation protocols while identifying critical open challenges in reward design, scalability, and cost-performance optimization. We further curate essential benchmarks, datasets, and metrics to facilitate rigorous assessment of post-training effectiveness. This survey aims to provide researchers and practitioners with a unified framework for advancing Video-LMM capabilities. Additional resources and updates are maintained at: https://github.com/yunlong10/Awesome-Video-LMM-Post-Training

Inter-frame modeling is pivotal in generating intermediate frames for video frame interpolation (VFI). Current approaches predominantly rely on convolution or attention-based models, which often either lack sufficient receptive fields or entail significant computational overheads. Recently, Selective State Space Models (S6) have emerged, tailored specifically for long sequence modeling, offering both linear complexity and data-dependent modeling capabilities. In this paper, we propose VFIMamba, a novel frame interpolation method for efficient and dynamic inter-frame modeling by harnessing the S6 model. Our approach introduces the Mixed-SSM Block (MSB), which initially rearranges tokens from adjacent frames in an interleaved fashion and subsequently applies multi-directional S6 modeling. This design facilitates the efficient transmission of information across frames while upholding linear complexity. Furthermore, we introduce a novel curriculum learning strategy that progressively cultivates proficiency in modeling inter-frame dynamics across varying motion magnitudes, fully unleashing the potential of the S6 model. Experimental findings showcase that our method attains state-of-the-art performance across diverse benchmarks, particularly excelling in high-resolution scenarios. In particular, on the X-TEST dataset, VFIMamba demonstrates a noteworthy improvement of 0.80 dB for 4K frames and 0.96 dB for 2K frames.

Vision-language models are integral to computer vision research, yet many high-performing models remain closed-source, obscuring their data, design and training recipe. The research community has responded by using distillation from black-box models to label training data, achieving strong benchmark results, at the cost of measurable scientific progress. However, without knowing the details of the teacher model and its data sources, scientific progress remains difficult to measure. In this paper, we study building a Perception Language Model (PLM) in a fully open and reproducible framework for transparent research in image and video understanding. We analyze standard training pipelines without distillation from proprietary models and explore large-scale synthetic data to identify critical data gaps, particularly in detailed video understanding. To bridge these gaps, we release 2.8M human-labeled instances of fine-grained video question-answer pairs and spatio-temporally grounded video captions. Additionally, we introduce PLM-VideoBench, a suite for evaluating challenging video understanding tasks focusing on the ability to reason about "what", "where", "when", and "how" of a video. We make our work fully reproducible by providing data, training recipes, code & models.

We introduce Human-like Video Models (HVM-1), large-scale video models pretrained with nearly 5000 hours of curated human-like video data (mostly egocentric, temporally extended, continuous video recordings), using the spatiotemporal masked autoencoder (ST-MAE) algorithm. We release two 633M parameter models trained at spatial resolutions of 224x224 and 448x448 pixels. We evaluate the performance of these models in downstream few-shot video and image recognition tasks and compare them against a model pretrained with 1330 hours of short action-oriented video clips from YouTube (Kinetics-700). HVM-1 models perform competitively against the Kinetics-700 pretrained model in downstream evaluations despite substantial qualitative differences between the spatiotemporal characteristics of the corresponding pretraining datasets. HVM-1 models also learn more accurate and more robust object representations compared to models pretrained with the image-based MAE algorithm on the same data, demonstrating the potential benefits of learning to predict temporal regularities in natural videos for learning better object representations.

Long video understanding is still challenging for recent Large Video-Language Models (LVLMs) due to the conflict between long-form temporal understanding and detailed spatial perception. LVLMs with a uniform frame sampling mechanism, which samples frames with an equal frame size and fixed sampling rate, inevitably sacrifice either temporal clues or spatial details, resulting in suboptimal solutions. To mitigate this dilemma, we propose LOVE-R1, a model that can adaptively zoom in on a video clip. The model is first provided with densely sampled frames but in a small resolution. If some spatial details are needed, the model can zoom in on a clip of interest with a large frame resolution based on its reasoning until key visual information is obtained. The whole process is implemented as a multi-step reasoning process. To train the reasoning ability, we first finetune the model on our collected 38k high-quality CoT data and enhance it with decoupled reinforcement finetuning. As outcome rewards can not provide fine-grained process supervision, we decouple multi-step reasoning into multiple single-step reasoning and optimize the internal zoom-in ability explicitly. Experiments on long video understanding benchmarks show that our model with the slow-fast adaptive frame sampling mechanism achieves a great trade-off between sampling density and frame resolutions, and LOVE-R1 outperforms our baseline Qwen2.5-VL by an average of 3.1% points across 4 common long video understanding benchmarks.

Recent advances in diffusion-based generative modeling have led to the development of text-to-video (T2V) models that can generate high-quality videos conditioned on a text prompt. Most of these T2V models often produce single-scene video clips that depict an entity performing a particular action (e.g., `a red panda climbing a tree'). However, it is pertinent to generate multi-scene videos since they are ubiquitous in the real-world (e.g., `a red panda climbing a tree' followed by `the red panda sleeps on the top of the tree'). To generate multi-scene videos from the pretrained T2V model, we introduce Time-Aligned Captions (TALC) framework. Specifically, we enhance the text-conditioning mechanism in the T2V architecture to recognize the temporal alignment between the video scenes and scene descriptions. For instance, we condition the visual features of the earlier and later scenes of the generated video with the representations of the first scene description (e.g., `a red panda climbing a tree') and second scene description (e.g., `the red panda sleeps on the top of the tree'), respectively. As a result, we show that the T2V model can generate multi-scene videos that adhere to the multi-scene text descriptions and be visually consistent (e.g., entity and background). Further, we finetune the pretrained T2V model with multi-scene video-text data using the TALC framework. We show that the TALC-finetuned model outperforms the baseline methods by 15.5 points in the overall score, which averages visual consistency and text adherence using human evaluation. The project website is https://talc-mst2v.github.io/.

This paper introduces a novel method for self-supervised video representation learning via feature prediction. In contrast to the previous methods that focus on future feature prediction, we argue that a supervisory signal arising from unobserved past frames is complementary to one that originates from the future frames. The rationale behind our method is to encourage the network to explore the temporal structure of videos by distinguishing between future and past given present observations. We train our model in a contrastive learning framework, where joint encoding of future and past provides us with a comprehensive set of temporal hard negatives via swapping. We empirically show that utilizing both signals enriches the learned representations for the downstream task of action recognition. It outperforms independent prediction of future and past.

Large Multimodal Models (LMMs) have demonstrated impressive performance in short video understanding tasks but face great challenges when applied to long video understanding. In contrast, Large Language Models (LLMs) exhibit outstanding capabilities in modeling long texts. Existing work attempts to address this issue by introducing long video-text pairs during training. However, these approaches require substantial computational and data resources. In this paper, we tackle the challenge of long video understanding from the perspective of context windows, aiming to apply LMMs to long video tasks without retraining on long video datasets. We first conduct an in-depth analysis of why pretrained LMMs struggle to understand lengthy video content, identifying that discrepancies between visual and language modalities lead to different context windows for visual and language tokens, making it difficult to directly extend the visual tokens to match the language context window. Based on this, we propose to adapt LMMs for long video understanding tasks by extending the visual context window, eliminating the need for retraining on large scalelong video datasets. To further mitigate the significant memory consumption caused by long sequences, we introduce a progressive pooling inference strategy that selectively adjusts the spatial resolution of frame embeddings, reducing the number of visual tokens while retaining important spatial information. Across multiple long video understanding benchmarks, our method consistently improves the performance as the number of video frames increases. On the MLVU benchmark, our method outperforms GPT-4o, even though our model size is only 7B. Additionally, in the 256-frame setting, our method reduces memory usage by approximately 45% compared to the baseline, without introducing any performance loss.

Standard discrete diffusion models treat all unobserved states identically by mapping them to an absorbing [MASK] token. This creates an 'information void' where semantic information that could be inferred from unmasked tokens is lost between denoising steps. We introduce Continuously Augmented Discrete Diffusion (CADD), a framework that augments the discrete state space with a paired diffusion in a continuous latent space. This yields graded, gradually corrupted states in which masked tokens are represented by noisy yet informative latent vectors rather than collapsed 'information voids'. At each reverse step, CADD may leverage the continuous latent as a semantic hint to guide discrete denoising. The design is clean and compatible with existing discrete diffusion training. At sampling time, the strength and choice of estimator for the continuous latent vector enables a controlled trade-off between mode-coverage (generating diverse outputs) and mode-seeking (generating contextually precise outputs) behaviors. Empirically, we demonstrate CADD improves generative quality over mask-based diffusion across text generation, image synthesis, and code modeling, with consistent gains on both qualitative and quantitative metrics against strong discrete baselines.

Previous works on video object segmentation (VOS) are trained on densely annotated videos. Nevertheless, acquiring annotations in pixel level is expensive and time-consuming. In this work, we demonstrate the feasibility of training a satisfactory VOS model on sparsely annotated videos-we merely require two labeled frames per training video while the performance is sustained. We term this novel training paradigm as two-shot video object segmentation, or two-shot VOS for short. The underlying idea is to generate pseudo labels for unlabeled frames during training and to optimize the model on the combination of labeled and pseudo-labeled data. Our approach is extremely simple and can be applied to a majority of existing frameworks. We first pre-train a VOS model on sparsely annotated videos in a semi-supervised manner, with the first frame always being a labeled one. Then, we adopt the pre-trained VOS model to generate pseudo labels for all unlabeled frames, which are subsequently stored in a pseudo-label bank. Finally, we retrain a VOS model on both labeled and pseudo-labeled data without any restrictions on the first frame. For the first time, we present a general way to train VOS models on two-shot VOS datasets. By using 7.3% and 2.9% labeled data of YouTube-VOS and DAVIS benchmarks, our approach achieves comparable results in contrast to the counterparts trained on fully labeled set. Code and models are available at https://github.com/yk-pku/Two-shot-Video-Object-Segmentation.

Understanding interior scenes has attracted enormous interest in computer vision community. However, few works focus on the understanding of furniture within the scenes and a large-scale dataset is also lacked to advance the field. In this paper, we first fill the gap by presenting DeepFurniture, a richly annotated large indoor scene dataset, including 24k indoor images, 170k furniture instances and 20k unique furniture identities. On the dataset, we introduce a new benchmark, named furniture set retrieval. Given an indoor photo as input, the task requires to detect all the furniture instances and search a matched set of furniture identities. To address this challenging task, we propose a feature and context embedding based framework. It contains 3 major contributions: (1) An improved Mask-RCNN model with an additional mask-based classifier is introduced for better utilizing the mask information to relieve the occlusion problems in furniture detection context. (2) A multi-task style Siamese network is proposed to train the feature embedding model for retrieval, which is composed of a classification subnet supervised by self-clustered pseudo attributes and a verification subnet to estimate whether the input pair is matched. (3) In order to model the relationship of the furniture entities in an interior design, a context embedding model is employed to re-rank the retrieval results. Extensive experiments demonstrate the effectiveness of each module and the overall system.

Existing work has observed that current text-to-image systems do not accurately reflect explicit spatial relations between objects such as 'left of' or 'below'. We hypothesize that this is because explicit spatial relations rarely appear in the image captions used to train these models. We propose an automatic method that, given existing images, generates synthetic captions that contain 14 explicit spatial relations. We introduce the Spatial Relation for Generation (SR4G) dataset, which contains 9.9 millions image-caption pairs for training, and more than 60 thousand captions for evaluation. In order to test generalization we also provide an 'unseen' split, where the set of objects in the train and test captions are disjoint. SR4G is the first dataset that can be used to spatially fine-tune text-to-image systems. We show that fine-tuning two different Stable Diffusion models (denoted as SD_{SR4G}) yields up to 9 points improvements in the VISOR metric. The improvement holds in the 'unseen' split, showing that SD_{SR4G} is able to generalize to unseen objects. SD_{SR4G} improves the state-of-the-art with fewer parameters, and avoids complex architectures. Our analysis shows that improvement is consistent for all relations. The dataset and the code will be publicly available.

Text-to-video (T2V) generation has made significant strides with diffusion models. However, existing methods still struggle with accurately binding attributes, determining spatial relationships, and capturing complex action interactions between multiple subjects. To address these limitations, we propose MagicComp, a training-free method that enhances compositional T2V generation through dual-phase refinement. Specifically, (1) During the Conditioning Stage: We introduce the Semantic Anchor Disambiguation to reinforces subject-specific semantics and resolve inter-subject ambiguity by progressively injecting the directional vectors of semantic anchors into original text embedding; (2) During the Denoising Stage: We propose Dynamic Layout Fusion Attention, which integrates grounding priors and model-adaptive spatial perception to flexibly bind subjects to their spatiotemporal regions through masked attention modulation. Furthermore, MagicComp is a model-agnostic and versatile approach, which can be seamlessly integrated into existing T2V architectures. Extensive experiments on T2V-CompBench and VBench demonstrate that MagicComp outperforms state-of-the-art methods, highlighting its potential for applications such as complex prompt-based and trajectory-controllable video generation. Project page: https://hong-yu-zhang.github.io/MagicComp-Page/.

Video Joint Embedding Predictive Architectures (V-JEPA) learn generalizable off-the-shelf video representation by predicting masked regions in latent space with an exponential moving average (EMA)-updated teacher. While EMA prevents representation collapse, it complicates scalable model selection and couples teacher and student architectures. We revisit masked-latent prediction and show that a frozen teacher suffices. Concretely, we (i) train a target encoder with a simple pixel-reconstruction objective under V-JEPA masking, then (ii) freeze it and train a student to predict the teacher's latents on masked regions. This leads to a two-stage, unregularized scheme that we refer to as SALT (Static-teacher Asymmetric Latent Training). SALT decouples optimization into pixel reconstruction (teacher) and masked latent prediction (student), increasing transparency, efficiency, and scalability while preserving the ability of representation to generalize under frozen evaluation. Empirically, our student models outperform recently proposed V-JEPA 2 encoders under frozen backbone evaluation across diverse benchmarks. They are also more compute-optimal: at matched pretraining FLOPs, our method achieves higher probing accuracy, and its scaling curves dominate V-JEPA's accuracy-FLOPs Pareto frontier. Finally, we find that student quality is remarkably robust to teacher quality: high-performing students emerge even with small, sub-optimal teachers. This points to a compute budget allocation that should overwhelmingly favor the student. These results position SALT as a simple, scalable, and compute-efficient alternative to EMA-based self-distillation for video representation learning.

Retinal foundation models aim to learn generalizable representations from diverse retinal images, facilitating label-efficient model adaptation across various ophthalmic tasks. Despite their success, current retinal foundation models are generally restricted to a single imaging modality, such as Color Fundus Photography (CFP) or Optical Coherence Tomography (OCT), limiting their versatility. Moreover, these models may struggle to fully leverage expert annotations and overlook the valuable domain knowledge essential for domain-specific representation learning. To overcome these limitations, we introduce UrFound, a retinal foundation model designed to learn universal representations from both multimodal retinal images and domain knowledge. UrFound is equipped with a modality-agnostic image encoder and accepts either CFP or OCT images as inputs. To integrate domain knowledge into representation learning, we encode expert annotation in text supervision and propose a knowledge-guided masked modeling strategy for model pre-training. It involves reconstructing randomly masked patches of retinal images while predicting masked text tokens conditioned on the corresponding retinal image. This approach aligns multimodal images and textual expert annotations within a unified latent space, facilitating generalizable and domain-specific representation learning. Experimental results demonstrate that UrFound exhibits strong generalization ability and data efficiency when adapting to various tasks in retinal image analysis. By training on ~180k retinal images, UrFound significantly outperforms the state-of-the-art retinal foundation model trained on up to 1.6 million unlabelled images across 8 public retinal datasets. Our code and data are available at https://github.com/yukkai/UrFound.

People say, "A picture is worth a thousand words". Then how can we get the rich information out of the image? We argue that by using visual clues to bridge large pretrained vision foundation models and language models, we can do so without any extra cross-modal training. Thanks to the strong zero-shot capability of foundation models, we start by constructing a rich semantic representation of the image (e.g., image tags, object attributes / locations, captions) as a structured textual prompt, called visual clues, using a vision foundation model. Based on visual clues, we use large language model to produce a series of comprehensive descriptions for the visual content, which is then verified by the vision model again to select the candidate that aligns best with the image. We evaluate the quality of generated descriptions by quantitative and qualitative measurement. The results demonstrate the effectiveness of such a structured semantic representation.

We present the TinyLLaVA-Video, a video understanding model with parameters not exceeding 4B that processes video sequences in a simple manner, without the need for complex architectures, supporting both fps sampling and uniform frame sampling. Our model is characterized by modularity and scalability, allowing training and inference with limited computational resources and enabling users to replace components based on their needs. We validate the effectiveness of this framework through experiments, the best model achieving performance comparable to certain existing 7B models on multiple video understanding benchmarks. The code and training recipes are fully open source, with all components and training data publicly available. We hope this work can serve as a baseline for practitioners exploring small-scale multimodal models for video understanding. It is available at https://github.com/ZhangXJ199/TinyLLaVA-Video.

Learning a good representation for space-time correspondence is the key for various computer vision tasks, including tracking object bounding boxes and performing video object pixel segmentation. To learn generalizable representation for correspondence in large-scale, a variety of self-supervised pretext tasks are proposed to explicitly perform object-level or patch-level similarity learning. Instead of following the previous literature, we propose to learn correspondence using Video Frame-level Similarity (VFS) learning, i.e, simply learning from comparing video frames. Our work is inspired by the recent success in image-level contrastive learning and similarity learning for visual recognition. Our hypothesis is that if the representation is good for recognition, it requires the convolutional features to find correspondence between similar objects or parts. Our experiments show surprising results that VFS surpasses state-of-the-art self-supervised approaches for both OTB visual object tracking and DAVIS video object segmentation. We perform detailed analysis on what matters in VFS and reveals new properties on image and frame level similarity learning. Project page with code is available at https://jerryxu.net/VFS

Text-to-image diffusion models have an unprecedented ability to generate diverse and high-quality images. However, they often struggle to faithfully capture the intended semantics of complex input prompts that include multiple subjects. Recently, numerous layout-to-image extensions have been introduced to improve user control, aiming to localize subjects represented by specific tokens. Yet, these methods often produce semantically inaccurate images, especially when dealing with multiple semantically or visually similar subjects. In this work, we study and analyze the causes of these limitations. Our exploration reveals that the primary issue stems from inadvertent semantic leakage between subjects in the denoising process. This leakage is attributed to the diffusion model's attention layers, which tend to blend the visual features of different subjects. To address these issues, we introduce Bounded Attention, a training-free method for bounding the information flow in the sampling process. Bounded Attention prevents detrimental leakage among subjects and enables guiding the generation to promote each subject's individuality, even with complex multi-subject conditioning. Through extensive experimentation, we demonstrate that our method empowers the generation of multiple subjects that better align with given prompts and layouts.

In this paper, we introduce a novel visual representation learning which relies on a handful of adaptively learned tokens, and which is applicable to both image and video understanding tasks. Instead of relying on hand-designed splitting strategies to obtain visual tokens and processing a large number of densely sampled patches for attention, our approach learns to mine important tokens in visual data. This results in efficiently and effectively finding a few important visual tokens and enables modeling of pairwise attention between such tokens, over a longer temporal horizon for videos, or the spatial content in images. Our experiments demonstrate strong performance on several challenging benchmarks for both image and video recognition tasks. Importantly, due to our tokens being adaptive, we accomplish competitive results at significantly reduced compute amount. We obtain comparable results to the state-of-the-arts on ImageNet while being computationally more efficient. We also confirm the effectiveness of the approach on multiple video datasets, including Kinetics-400, Kinetics-600, Charades, and AViD. The code is available at: https://github.com/google-research/scenic/tree/main/scenic/projects/token_learner

Zero-shot, training-free, image-based text-to-video generation is an emerging area that aims to generate videos using existing image-based diffusion models. Current methods in this space require specific architectural changes to image generation models, which limit their adaptability and scalability. In contrast to such methods, we provide a model-agnostic approach. We use intersections in diffusion trajectories, working only with the latent values. We could not obtain localized frame-wise coherence and diversity using only the intersection of trajectories. Thus, we instead use a grid-based approach. An in-context trained LLM is used to generate coherent frame-wise prompts; another is used to identify differences between frames. Based on these, we obtain a CLIP-based attention mask that controls the timing of switching the prompts for each grid cell. Earlier switching results in higher variance, while later switching results in more coherence. Therefore, our approach can ensure appropriate control between coherence and variance for the frames. Our approach results in state-of-the-art performance while being more flexible when working with diverse image-generation models. The empirical analysis using quantitative metrics and user studies confirms our model's superior temporal consistency, visual fidelity and user satisfaction, thus providing a novel way to obtain training-free, image-based text-to-video generation.

Driven by the success of Masked Language Modeling (MLM), the realm of self-supervised learning for computer vision has been invigorated by the central role of Masked Image Modeling (MIM) in driving recent breakthroughs. Notwithstanding the achievements of MIM across various downstream tasks, its overall efficiency is occasionally hampered by the lengthy duration of the pre-training phase. This paper presents a perspective that the optimization of masked tokens as a means of addressing the prevailing issue. Initially, we delve into an exploration of the inherent properties that a masked token ought to possess. Within the properties, we principally dedicated to articulating and emphasizing the `data singularity' attribute inherent in masked tokens. Through a comprehensive analysis of the heterogeneity between masked tokens and visible tokens within pre-trained models, we propose a novel approach termed masked token optimization (MTO), specifically designed to improve model efficiency through weight recalibration and the enhancement of the key property of masked tokens. The proposed method serves as an adaptable solution that seamlessly integrates into any MIM approach that leverages masked tokens. As a result, MTO achieves a considerable improvement in pre-training efficiency, resulting in an approximately 50% reduction in pre-training epochs required to attain converged performance of the recent approaches.

Text-conditioned diffusion models have emerged as a promising tool for neural video generation. However, current models still struggle with intricate spatiotemporal prompts and often generate restricted or incorrect motion (e.g., even lacking the ability to be prompted for objects moving from left to right). To address these limitations, we introduce LLM-grounded Video Diffusion (LVD). Instead of directly generating videos from the text inputs, LVD first leverages a large language model (LLM) to generate dynamic scene layouts based on the text inputs and subsequently uses the generated layouts to guide a diffusion model for video generation. We show that LLMs are able to understand complex spatiotemporal dynamics from text alone and generate layouts that align closely with both the prompts and the object motion patterns typically observed in the real world. We then propose to guide video diffusion models with these layouts by adjusting the attention maps. Our approach is training-free and can be integrated into any video diffusion model that admits classifier guidance. Our results demonstrate that LVD significantly outperforms its base video diffusion model and several strong baseline methods in faithfully generating videos with the desired attributes and motion patterns.

We propose a novel approach for few-shot talking-head synthesis. While recent works in neural talking heads have produced promising results, they can still produce images that do not preserve the identity of the subject in source images. We posit this is a result of the entangled representation of each subject in a single latent code that models 3D shape information, identity cues, colors, lighting and even background details. In contrast, we propose to factorize the representation of a subject into its spatial and style components. Our method generates a target frame in two steps. First, it predicts a dense spatial layout for the target image. Second, an image generator utilizes the predicted layout for spatial denormalization and synthesizes the target frame. We experimentally show that this disentangled representation leads to a significant improvement over previous methods, both quantitatively and qualitatively.

Visual understanding is inherently contextual -- what we focus on in an image depends on the task at hand. For instance, given an image of a person holding a bouquet of flowers, we may focus on either the person such as their clothing, or the type of flowers, depending on the context of interest. Yet, most existing image encoding paradigms represent an image as a fixed, generic feature vector, overlooking the potential needs of prioritizing varying visual information for different downstream use cases. In this work, we introduce FocalLens, a conditional visual encoding method that produces different representations for the same image based on the context of interest, expressed flexibly through natural language. We leverage vision instruction tuning data and contrastively finetune a pretrained vision encoder to take natural language instructions as additional inputs for producing conditional image representations. Extensive experiments validate that conditional image representation from FocalLens better pronounce the visual features of interest compared to generic features produced by standard vision encoders like CLIP. In addition, we show FocalLens further leads to performance improvements on a range of downstream tasks including image-image retrieval, image classification, and image-text retrieval, with an average gain of 5 and 10 points on the challenging SugarCrepe and MMVP-VLM benchmarks, respectively.

Despite their great success across various multimodal tasks, Large Vision-Language Models (LVLMs) are facing a prevalent problem with object hallucinations, where the generated textual responses are inconsistent with ground-truth objects in the given image. This paper investigates various LVLMs and pinpoints attention deficiency toward discriminative local image features as one root cause of object hallucinations. Specifically, LVLMs predominantly attend to prompt-independent global image features, while failing to capture prompt-relevant local features, consequently undermining the visual grounding capacity of LVLMs and leading to hallucinations. To this end, we propose Assembly of Global and Local Attention (AGLA), a training-free and plug-and-play approach that mitigates object hallucinations by exploring an ensemble of global features for response generation and local features for visual discrimination simultaneously. Our approach exhibits an image-prompt matching scheme that captures prompt-relevant local features from images, leading to an augmented view of the input image where prompt-relevant content is reserved while irrelevant distractions are masked. With the augmented view, a calibrated decoding distribution can be derived by integrating generative global features from the original image and discriminative local features from the augmented image. Extensive experiments show that AGLA consistently mitigates object hallucinations and enhances general perception capability for LVLMs across various discriminative and generative benchmarks. Our code will be released at https://github.com/Lackel/AGLA.

Diffusion model is a promising approach to image generation and has been employed for Pose-Guided Person Image Synthesis (PGPIS) with competitive performance. While existing methods simply align the person appearance to the target pose, they are prone to overfitting due to the lack of a high-level semantic understanding on the source person image. In this paper, we propose a novel Coarse-to-Fine Latent Diffusion (CFLD) method for PGPIS. In the absence of image-caption pairs and textual prompts, we develop a novel training paradigm purely based on images to control the generation process of the pre-trained text-to-image diffusion model. A perception-refined decoder is designed to progressively refine a set of learnable queries and extract semantic understanding of person images as a coarse-grained prompt. This allows for the decoupling of fine-grained appearance and pose information controls at different stages, and thus circumventing the potential overfitting problem. To generate more realistic texture details, a hybrid-granularity attention module is proposed to encode multi-scale fine-grained appearance features as bias terms to augment the coarse-grained prompt. Both quantitative and qualitative experimental results on the DeepFashion benchmark demonstrate the superiority of our method over the state of the arts for PGPIS. Code is available at https://github.com/YanzuoLu/CFLD.

This paper introduces VideoScan, an efficient vision-language model (VLM) inference framework designed for real-time video interaction that effectively comprehends and retains streamed video inputs while delivering rapid and accurate responses. A longstanding challenge in video understanding--particularly for long-term or real-time applications--stems from the substantial computational overhead caused by the extensive length of visual tokens. To address this, VideoScan employs a single semantic carrier token to represent each frame, progressively reducing computational and memory overhead during its two-phase inference process: prefilling and decoding. The embedding of the semantic carrier token is derived from an optimized aggregation of frame-level visual features, ensuring compact yet semantically rich representations. Critically, the corresponding key-value pairs are trained to retain contextual semantics from prior frames, enabling efficient memory management without sacrificing temporal coherence. During inference, the visual tokens of each frame are processed only once during the prefilling phase and subsequently discarded in the decoding stage, eliminating redundant computations. This design ensures efficient VLM inference even under stringent real-time constraints. Comprehensive experiments on diverse offline and online benchmarks demonstrate that LLaVA-Video, supported by our method, achieves up to sim 5times and 1.29times speedups compared to its original version and previous efficient streaming video understanding approaches, respectively. Crucially, these improvements are attained while maintaining competitive performance and ensuring stable GPU memory consumption (consistently sim 18GB, independent of video duration).

Existing video-language pre-training methods primarily focus on instance-level alignment between video clips and captions via global contrastive learning but neglect rich fine-grained local information in both videos and text, which is of importance to downstream tasks requiring temporal localization and semantic reasoning. A powerful model is expected to be capable of capturing region-object correspondences and recognizing scene changes in a video clip, reflecting spatial and temporal granularity, respectively. To strengthen model's understanding into such fine-grained details, we propose a simple yet effective video-language modeling framework, S-ViLM, by exploiting the intrinsic structures of these two modalities. It includes two novel designs, inter-clip spatial grounding and intra-clip temporal grouping, to promote learning region-object alignment and temporal-aware features, simultaneously. Comprehensive evaluations demonstrate that S-ViLM performs favorably against existing approaches in learning more expressive representations. Specifically, S-ViLM surpasses the state-of-the-art methods substantially on four representative downstream tasks, covering text-video retrieval, video question answering, video action recognition, and temporal action localization.

How multimodal large language models (MLLMs) perceive the visual world remains a mystery. To one extreme, object and relation modeling may be implicitly implemented with inductive biases, for example by treating objects as tokens. To the other extreme, empirical results reveal the surprising finding that simply performing visual captioning, which tends to ignore spatial configuration of the objects, serves as a strong baseline for video understanding. We aim to answer the question: how can objects help video-language understanding in MLLMs? We tackle the question from the object representation and adaptation perspectives. Specifically, we investigate the trade-off between representation expressiveness (e.g., distributed versus symbolic) and integration difficulty (e.g., data-efficiency when learning the adapters). Through extensive evaluations on five video question answering datasets, we confirm that explicit integration of object-centric representation remains necessary, and the symbolic objects can be most easily integrated while being performant for question answering. We hope our findings can encourage the community to explore the explicit integration of perception modules into MLLM design. Our code and models will be publicly released.

We propose a novel approach for captioning and object grounding in video, where the objects in the caption are grounded in the video via temporally dense bounding boxes. We introduce the following contributions. First, we present a large-scale automatic annotation method that aggregates captions grounded with bounding boxes across individual frames into temporally dense and consistent bounding box annotations. We apply this approach on the HowTo100M dataset to construct a large-scale pre-training dataset, named HowToGround1M. We also introduce a Grounded Video Caption Generation model, dubbed GROVE, and pre-train the model on HowToGround1M. Second, we introduce a new dataset, called iGround, of 3500 videos with manually annotated captions and dense spatio-temporally grounded bounding boxes. This allows us to measure progress on this challenging problem, as well as to fine-tune our model on this small-scale but high-quality data. Third, we demonstrate that our approach achieves state-of-the-art results on the proposed iGround dataset compared to a number of baselines, as well as on the VidSTG and ActivityNet-Entities datasets. We perform extensive ablations that demonstrate the importance of pre-training using our automatically annotated HowToGround1M dataset followed by fine-tuning on the manually annotated iGround dataset and validate the key technical contributions of our model.

In this paper, we address a challenging task, synchronous motion captioning, that aim to generate a language description synchronized with human motion sequences. This task pertains to numerous applications, such as aligned sign language transcription, unsupervised action segmentation and temporal grounding. Our method introduces mechanisms to control self- and cross-attention distributions of the Transformer, allowing interpretability and time-aligned text generation. We achieve this through masking strategies and structuring losses that push the model to maximize attention only on the most important frames contributing to the generation of a motion word. These constraints aim to prevent undesired mixing of information in attention maps and to provide a monotonic attention distribution across tokens. Thus, the cross attentions of tokens are used for progressive text generation in synchronization with human motion sequences. We demonstrate the superior performance of our approach through evaluation on the two available benchmark datasets, KIT-ML and HumanML3D. As visual evaluation is essential for this task, we provide a comprehensive set of animated visual illustrations in the code repository: https://github.com/rd20karim/Synch-Transformer.

Vision-Language Models (VLMs) have shown remarkable capabilities across diverse visual tasks, including image recognition, video understanding, and Visual Question Answering (VQA) when explicitly trained for these tasks. Despite these advances, we find that current VLMs lack a fundamental cognitive ability: learning to localize objects in a scene by taking into account the context. In this work, we focus on the task of few-shot personalized localization, where a model is given a small set of annotated images (in-context examples) -- each with a category label and bounding box -- and is tasked with localizing the same object type in a query image. To provoke personalized localization abilities in models, we present a data-centric solution that fine-tunes them using carefully curated data from video object tracking datasets. By leveraging sequences of frames tracking the same object across multiple shots, we simulate instruction-tuning dialogues that promote context awareness. To reinforce this, we introduce a novel regularization technique that replaces object labels with pseudo-names, ensuring the model relies on visual context rather than prior knowledge. Our method significantly enhances few-shot localization performance without sacrificing generalization, as demonstrated on several benchmarks tailored to personalized localization. This work is the first to explore and benchmark personalized few-shot localization for VLMs, laying a foundation for future research in context-driven vision-language applications. The code for our project is available at https://github.com/SivanDoveh/IPLoc

We propose a margin-based loss for vision-language model pretraining that encourages gradient-based explanations that are consistent with region-level annotations. We refer to this objective as Attention Mask Consistency (AMC) and demonstrate that it produces superior visual grounding performance compared to models that rely instead on region-level annotations for explicitly training an object detector such as Faster R-CNN. AMC works by encouraging gradient-based explanation masks that focus their attention scores mostly within annotated regions of interest for images that contain such annotations. Particularly, a model trained with AMC on top of standard vision-language modeling objectives obtains a state-of-the-art accuracy of 86.59% in the Flickr30k visual grounding benchmark, an absolute improvement of 5.48% when compared to the best previous model. Our approach also performs exceedingly well on established benchmarks for referring expression comprehension and offers the added benefit by design of gradient-based explanations that better align with human annotations.

Vision systems to see and reason about the compositional nature of visual scenes are fundamental to understanding our world. The complex relations between objects and their locations, ambiguities, and variations in the real-world environment can be better described in human language, naturally governed by grammatical rules and other modalities such as audio and depth. The models learned to bridge the gap between such modalities coupled with large-scale training data facilitate contextual reasoning, generalization, and prompt capabilities at test time. These models are referred to as foundational models. The output of such models can be modified through human-provided prompts without retraining, e.g., segmenting a particular object by providing a bounding box, having interactive dialogues by asking questions about an image or video scene or manipulating the robot's behavior through language instructions. In this survey, we provide a comprehensive review of such emerging foundational models, including typical architecture designs to combine different modalities (vision, text, audio, etc), training objectives (contrastive, generative), pre-training datasets, fine-tuning mechanisms, and the common prompting patterns; textual, visual, and heterogeneous. We discuss the open challenges and research directions for foundational models in computer vision, including difficulties in their evaluations and benchmarking, gaps in their real-world understanding, limitations of their contextual understanding, biases, vulnerability to adversarial attacks, and interpretability issues. We review recent developments in this field, covering a wide range of applications of foundation models systematically and comprehensively. A comprehensive list of foundational models studied in this work is available at https://github.com/awaisrauf/Awesome-CV-Foundational-Models.

Visual effects (VFX) are crucial to the expressive power of digital media, yet their creation remains a major challenge for generative AI. Prevailing methods often rely on the one-LoRA-per-effect paradigm, which is resource-intensive and fundamentally incapable of generalizing to unseen effects, thus limiting scalability and creation. To address this challenge, we introduce VFXMaster, the first unified, reference-based framework for VFX video generation. It recasts effect generation as an in-context learning task, enabling it to reproduce diverse dynamic effects from a reference video onto target content. In addition, it demonstrates remarkable generalization to unseen effect categories. Specifically, we design an in-context conditioning strategy that prompts the model with a reference example. An in-context attention mask is designed to precisely decouple and inject the essential effect attributes, allowing a single unified model to master the effect imitation without information leakage. In addition, we propose an efficient one-shot effect adaptation mechanism to boost generalization capability on tough unseen effects from a single user-provided video rapidly. Extensive experiments demonstrate that our method effectively imitates various categories of effect information and exhibits outstanding generalization to out-of-domain effects. To foster future research, we will release our code, models, and a comprehensive dataset to the community.

We present Pix2Gif, a motion-guided diffusion model for image-to-GIF (video) generation. We tackle this problem differently by formulating the task as an image translation problem steered by text and motion magnitude prompts, as shown in teaser fig. To ensure that the model adheres to motion guidance, we propose a new motion-guided warping module to spatially transform the features of the source image conditioned on the two types of prompts. Furthermore, we introduce a perceptual loss to ensure the transformed feature map remains within the same space as the target image, ensuring content consistency and coherence. In preparation for the model training, we meticulously curated data by extracting coherent image frames from the TGIF video-caption dataset, which provides rich information about the temporal changes of subjects. After pretraining, we apply our model in a zero-shot manner to a number of video datasets. Extensive qualitative and quantitative experiments demonstrate the effectiveness of our model -- it not only captures the semantic prompt from text but also the spatial ones from motion guidance. We train all our models using a single node of 16xV100 GPUs. Code, dataset and models are made public at: https://hiteshk03.github.io/Pix2Gif/.

Building on the advances of language models, Large Multimodal Models (LMMs) have contributed significant improvements in video understanding. While the current video LMMs utilize advanced Large Language Models (LLMs), they rely on either image or video encoders to process visual inputs, each of which has its own limitations. Image encoders excel at capturing rich spatial details from frame sequences but lack explicit temporal context, which can be important in videos with intricate action sequences. On the other hand, video encoders provide temporal context but are often limited by computational constraints that lead to processing only sparse frames at lower resolutions, resulting in reduced contextual and spatial understanding. To this end, we introduce VideoGPT+, which combines the complementary benefits of the image encoder (for detailed spatial understanding) and the video encoder (for global temporal context modeling). The model processes videos by dividing them into smaller segments and applies an adaptive pooling strategy on features extracted by both image and video encoders. Our architecture showcases improved performance across multiple video benchmarks, including VCGBench, MVBench and Zero-shot question-answering. Further, we develop 112K video-instruction set using a novel semi-automatic annotation pipeline which further improves the model performance. Additionally, to comprehensively evaluate video LMMs, we present VCGBench-Diverse, covering 18 broad video categories such as lifestyle, sports, science, gaming, and surveillance videos. This benchmark with 4,354 question-answer pairs evaluates the generalization of existing LMMs on dense video captioning, spatial and temporal understanding, and complex reasoning, ensuring comprehensive assessment across diverse video types and dynamics. Code: https://github.com/mbzuai-oryx/VideoGPT-plus.

Generative Large Multimodal Models (LMMs) like LLaVA and Qwen-VL excel at a wide variety of vision-language (VL) tasks such as image captioning or visual question answering. Despite strong performance, LMMs are not directly suited for foundational discriminative vision-language tasks (i.e., tasks requiring discrete label predictions) such as image classification and multiple-choice VQA. One key challenge in utilizing LMMs for discriminative tasks is the extraction of useful features from generative models. To overcome this issue, we propose an approach for finding features in the model's latent space to more effectively leverage LMMs for discriminative tasks. Toward this end, we present Sparse Attention Vectors (SAVs) -- a finetuning-free method that leverages sparse attention head activations (fewer than 1\% of the heads) in LMMs as strong features for VL tasks. With only few-shot examples, SAVs demonstrate state-of-the-art performance compared to a variety of few-shot and finetuned baselines on a collection of discriminative tasks. Our experiments also imply that SAVs can scale in performance with additional examples and generalize to similar tasks, establishing SAVs as both effective and robust multimodal feature representations.

For Large Vision-Language Models (LVLMs), scaling the model can effectively improve performance. However, expanding model parameters significantly increases the training and inferring costs, as all model parameters are activated for each token in the calculation. In this work, we propose a novel training strategy MoE-tuning for LVLMs, which can constructing a sparse model with an outrageous number of parameter but a constant computational cost, and effectively addresses the performance degradation typically associated with multi-modal learning and model sparsity. Furthermore, we present the MoE-LLaVA framework, a MoE-based sparse LVLM architecture. This framework uniquely activates only the top-k experts through routers during deployment, keeping the remaining experts inactive. Our extensive experiments highlight the excellent capabilities of MoE-LLaVA in visual understanding and its potential to reduce hallucinations in model outputs. Remarkably, with just 3 billion sparsely activated parameters, MoE-LLaVA demonstrates performance comparable to the LLaVA-1.5-7B on various visual understanding datasets and even surpasses the LLaVA-1.5-13B in object hallucination benchmarks. Through MoE-LLaVA, we aim to establish a baseline for sparse LVLMs and provide valuable insights for future research in developing more efficient and effective multi-modal learning systems. Code is released at https://github.com/PKU-YuanGroup/MoE-LLaVA.

Large video-language models (VLMs) have demonstrated promising progress in various video understanding tasks. However, their effectiveness in long-form video analysis is constrained by limited context windows. Traditional approaches, such as uniform frame sampling, often inevitably allocate resources to irrelevant content, diminishing their effectiveness in real-world scenarios. In this paper, we introduce BOLT, a method to BOost Large VLMs without additional Training through a comprehensive study of frame selection strategies. First, to enable a more realistic evaluation of VLMs in long-form video understanding, we propose a multi-source retrieval evaluation setting. Our findings reveal that uniform sampling performs poorly in noisy contexts, underscoring the importance of selecting the right frames. Second, we explore several frame selection strategies based on query-frame similarity and analyze their effectiveness at inference time. Our results show that inverse transform sampling yields the most significant performance improvement, increasing accuracy on the Video-MME benchmark from 53.8% to 56.1% and MLVU benchmark from 58.9% to 63.4%. Our code is available at https://github.com/sming256/BOLT.

The ability to integrate context, including perceptual and temporal cues, plays a pivotal role in grounding the meaning of a linguistic utterance. In order to measure to what extent current vision-and-language models master this ability, we devise a new multimodal challenge, Image Retrieval from Contextual Descriptions (ImageCoDe). In particular, models are tasked with retrieving the correct image from a set of 10 minimally contrastive candidates based on a contextual description. As such, each description contains only the details that help distinguish between images. Because of this, descriptions tend to be complex in terms of syntax and discourse and require drawing pragmatic inferences. Images are sourced from both static pictures and video frames. We benchmark several state-of-the-art models, including both cross-encoders such as ViLBERT and bi-encoders such as CLIP, on ImageCoDe. Our results reveal that these models dramatically lag behind human performance: the best variant achieves an accuracy of 20.9 on video frames and 59.4 on static pictures, compared with 90.8 in humans. Furthermore, we experiment with new model variants that are better equipped to incorporate visual and temporal context into their representations, which achieve modest gains. Our hope is that ImageCoDE will foster progress in grounded language understanding by encouraging models to focus on fine-grained visual differences.

Motivated by the hypothesis that neural network representations encode abstract, interpretable features as linearly accessible, approximately orthogonal directions, sparse autoencoders (SAEs) have become a popular tool in interpretability. However, recent work has demonstrated phenomenology of model representations that lies outside the scope of this hypothesis, showing signatures of hierarchical, nonlinear, and multi-dimensional features. This raises the question: do SAEs represent features that possess structure at odds with their motivating hypothesis? If not, does avoiding this mismatch help identify said features and gain further insights into neural network representations? To answer these questions, we take a construction-based approach and re-contextualize the popular matching pursuits (MP) algorithm from sparse coding to design MP-SAE -- an SAE that unrolls its encoder into a sequence of residual-guided steps, allowing it to capture hierarchical and nonlinearly accessible features. Comparing this architecture with existing SAEs on a mixture of synthetic and natural data settings, we show: (i) hierarchical concepts induce conditionally orthogonal features, which existing SAEs are unable to faithfully capture, and (ii) the nonlinear encoding step of MP-SAE recovers highly meaningful features, helping us unravel shared structure in the seemingly dichotomous representation spaces of different modalities in a vision-language model, hence demonstrating the assumption that useful features are solely linearly accessible is insufficient. We also show that the sequential encoder principle of MP-SAE affords an additional benefit of adaptive sparsity at inference time, which may be of independent interest. Overall, we argue our results provide credence to the idea that interpretability should begin with the phenomenology of representations, with methods emerging from assumptions that fit it.

This paper shows that masked autoencoders (MAE) are scalable self-supervised learners for computer vision. Our MAE approach is simple: we mask random patches of the input image and reconstruct the missing pixels. It is based on two core designs. First, we develop an asymmetric encoder-decoder architecture, with an encoder that operates only on the visible subset of patches (without mask tokens), along with a lightweight decoder that reconstructs the original image from the latent representation and mask tokens. Second, we find that masking a high proportion of the input image, e.g., 75%, yields a nontrivial and meaningful self-supervisory task. Coupling these two designs enables us to train large models efficiently and effectively: we accelerate training (by 3x or more) and improve accuracy. Our scalable approach allows for learning high-capacity models that generalize well: e.g., a vanilla ViT-Huge model achieves the best accuracy (87.8%) among methods that use only ImageNet-1K data. Transfer performance in downstream tasks outperforms supervised pre-training and shows promising scaling behavior.

Zero-shot learning has gained popularity due to its potential to scale recognition models without requiring additional training data. This is usually achieved by associating categories with their semantic information like attributes. However, we believe that the potential offered by this paradigm is not yet fully exploited. In this work, we propose to utilize the structure of the space spanned by the attributes using a set of relations. We devise objective functions to preserve these relations in the embedding space, thereby inducing semanticity to the embedding space. Through extensive experimental evaluation on five benchmark datasets, we demonstrate that inducing semanticity to the embedding space is beneficial for zero-shot learning. The proposed approach outperforms the state-of-the-art on the standard zero-shot setting as well as the more realistic generalized zero-shot setting. We also demonstrate how the proposed approach can be useful for making approximate semantic inferences about an image belonging to a category for which attribute information is not available.

Segmenting an object in a video presents significant challenges. Each pixel must be accurately labelled, and these labels must remain consistent across frames. The difficulty increases when the segmentation is with arbitrary granularity, meaning the number of segments can vary arbitrarily, and masks are defined based on only one or a few sample images. In this paper, we address this issue by employing a pre-trained text to image diffusion model supplemented with an additional tracking mechanism. We demonstrate that our approach can effectively manage various segmentation scenarios and outperforms state-of-the-art alternatives.

Foundation models like the Segment Anything Model (SAM) have significantly advanced promptable image segmentation in computer vision. However, extending these capabilities to videos presents substantial challenges, particularly in ensuring precise and temporally consistent mask propagation in dynamic scenes. SAM 2 attempts to address this by training a model on massive image and video data from scratch to learn complex spatiotemporal associations, resulting in huge training costs that hinder research and practical deployment. In this paper, we introduce SAM-I2V, an effective image-to-video upgradation method for cultivating a promptable video segmentation (PVS) model. Our approach strategically upgrades the pre-trained SAM to support PVS, significantly reducing training complexity and resource requirements. To achieve this, we introduce three key innovations: (i) an image-to-video feature extraction upgrader built upon SAM's static image encoder to enable spatiotemporal video perception, (ii) a memory filtering strategy that selects the most relevant past frames for more effective utilization of historical information, and (iii) a memory-as-prompt mechanism leveraging object memory to ensure temporally consistent mask propagation in dynamic scenes. Comprehensive experiments demonstrate that our method achieves over 90% of SAM 2's performance while using only 0.2% of its training cost. Our work presents a resource-efficient pathway to PVS, lowering barriers for further research in PVS model design and enabling broader applications and advancements in the field. Code and model are available at: https://github.com/showlab/SAM-I2V.