Width and Depth Limits Commute in Residual Networks
Paper
•
2302.00453
•
Published
Soufiane Hayou
soufiane001
AI & ML interests
AI Scientist at UC Berkeley
Organizations
None yet
authored
12
papers
over 1 year ago
Tensor Programs VI: Feature Learning in Infinite-Depth Neural Networks
Paper
•
2310.02244
•
Published
Leave-one-out Distinguishability in Machine Learning
Paper
•
2309.17310
•
Published
On the Impact of the Activation Function on Deep Neural Networks Training
Paper
•
1902.06853
•
Published
Commutative Width and Depth Scaling in Deep Neural Networks
Paper
•
2310.01683
•
Published
Data pruning and neural scaling laws: fundamental limitations of score-based algorithms
Paper
•
2302.06960
•
Published
On the infinite-depth limit of finite-width neural networks
Paper
•
2210.00688
•
Published
From Optimization Dynamics to Generalization Bounds via Łojasiewicz Gradient Inequality
Paper
•
2202.10670
•
Published
Feature Learning and Signal Propagation in Deep Neural Networks
Paper
•
2110.11749
•
Published
Stable ResNet
Paper
•
2010.12859
•
Published
Robust Pruning at Initialization
Paper
•
2002.08797
•
Published
LoRA+: Efficient Low Rank Adaptation of Large Models
Paper
•
2402.12354
•
Published
•
6