Theory

How and When Random Feedback Works: A Case Study of Low-Rank Matrix Factorization AISTATS 2022 How Much More Data Do I Need? Estimating Requirements for Downstream Tasks CVPR 2022 How Powerful are K-hop Message Passing Graph Neural Networks NeurIPS 2022 How Sampling Impacts the Robustness of Stochastic Neural Networks NeurIPS 2022 Identifying good directions to escape the NTK regime and efficiently learn low-degree plus sparse polynomials NeurIPS 2022 Imbalance Trouble: Revisiting Neural-Collapse Geometry NeurIPS 2022 Implicit Bias of Linear Equivariant Networks ICML 2022 Implicit Bias of the Step Size in Linear Diagonal Neural Networks ICML 2022 Implicit Regularization or Implicit Conditioning? Exact Risk Trajectories of SGD in High Dimensions NeurIPS 2022 Implicit Regularization with Polynomial Growth in Deep Tensor Factorization ICML 2022 Improved Bounds on Neural Complexity for Representing Piecewise Linear Functions NeurIPS 2022 Individual Fairness Guarantees for Neural Networks IJCAI 2022 Inductive Bias of Multi-Channel Linear Convolutional Networks with Bounded Weight Norm COLT 2022 Informed Learning by Wide Neural Networks: Convergence, Generalization and Sampling Complexity ICML 2022 Intrinsic Dimension Estimation Using Wasserstein Distance JMLR 2022 Invariance-Aware Randomized Smoothing Certificates NeurIPS 2022 In What Ways Are Deep Neural Networks Invariant and How Should We Measure This? NeurIPS 2022 KAM Theory Meets Statistical Learning Theory: Hamiltonian Neural Networks with Non-zero Training Loss AAAI 2022 Label noise (stochastic) gradient descent implicitly solves the Lasso for quadratic parametrisation COLT 2022 Language Model Decomposition: Quantifying the Dependency and Correlation of Language Models EMNLP 2022 Learning dynamics of deep linear networks with multiple pathways NeurIPS 2022 Learning Rates as a Function of Batch Size: A Random Matrix Theory Approach to Neural Network Training JMLR 2022 Learning sparse features can lead to overfitting in neural networks NeurIPS 2022 Learning to Reason with Neural Networks: Generalization, Unseen Data and Boolean Measures NeurIPS 2022 Listen to Interpret: Post-hoc Interpretability for Audio Networks with NMF NeurIPS 2022