MergeMix: A Unified Augmentation Paradigm for Visual and Multi-Modal Understanding Paper • 2510.23479 • Published 2 days ago • 11
Met^2Net: A Decoupled Two-Stage Spatio-Temporal Forecasting Model for Complex Meteorological Systems Paper • 2507.17189 • Published Jul 23 • 12
Rep-MTL: Unleashing the Power of Representation-level Task Saliency for Multi-Task Learning Paper • 2507.21049 • Published Jul 28 • 40
MergeVQ: A Unified Framework for Visual Generation and Representation with Disentangled Token Merging and Quantization Paper • 2504.00999 • Published Apr 1 • 93
Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning Paper • 2506.01939 • Published Jun 2 • 185
Taming LLMs by Scaling Learning Rates with Gradient Grouping Paper • 2506.01049 • Published Jun 1 • 38