Theory

A Quantitative Geometric Approach to Neural-Network Smoothness NeurIPS 2022 A Random Matrix Perspective on Mixtures of Nonlinearities in High Dimensions AISTATS 2022 Are All Losses Created Equal: A Neural Collapse Perspective NeurIPS 2022 A Reparametrization-Invariant Sharpness Measure Based on Information Geometry NeurIPS 2022 A Spectral Perspective of DNN Robustness to Label Noise AISTATS 2022 A Theoretical View on Sparsely Activated Networks NeurIPS 2022 A Unified Convergence Theorem for Stochastic Optimization Methods NeurIPS 2022 Automatic differentiation of nonsmooth iterative algorithms NeurIPS 2022 Automatic Differentiation of Programs with Discrete Randomness NeurIPS 2022 Benefits of Additive Noise in Composing Classes with Bounded Capacity NeurIPS 2022 Benefits of Overparameterized Convolutional Residual Networks: Function Approximation under Smoothness Constraint ICML 2022 Benign Overfitting in Two-layer Convolutional Neural Networks NeurIPS 2022 Benign Overfitting without Linearity: Neural Network Classifiers Trained by Gradient Descent for Noisy Linear Data COLT 2022 Benign Underfitting of Stochastic Gradient Descent NeurIPS 2022 Beyond accuracy: generalization properties of bio-plausible temporal credit assignment rules NeurIPS 2022 Beyond GNNs: An Efficient Architecture for Graph Problems AAAI 2022 Bounding the Width of Neural Networks via Coupled Initialization A Worst Case Analysis ICML 2022 Can Neural Nets Learn the Same Model Twice? Investigating Reproducibility and Double Descent From the Decision Boundary Perspective CVPR 2022 Can Push-forward Generative Models Fit Multimodal Distributions? NeurIPS 2022 CC-CERT: A Probabilistic Approach to Certify General Robustness of Neural Networks AAAI 2022 Chaotic Dynamics are Intrinsic to Neural Network Training with SGD NeurIPS 2022 Characterizing Datapoints via Second-Split Forgetting NeurIPS 2022 Convergence beyond the over-parameterized regime using Rayleigh quotients NeurIPS 2022 Convergence for score-based generative modeling with polynomial complexity NeurIPS 2022 Convergence Rates of Non-Convex Stochastic Gradient Descent Under a Generic Lojasiewicz Condition and Local Smoothness ICML 2022