Arya Mazumdar

On the Theory of Continual Learning with Gradient Descent for Neural Networks

Effectiveness of Local Steps on Heterogeneous Data: An Implicit Bias View

Exact Recovery of Sparse Binary Vectors from Generalized Linear Measurements

Sharper Guarantees for Learning Neural Network Classifiers with Gradient Methods

Generalization Bound of Gradient Flow through Training Trajectory and Data-dependent Kernel

Optimal Transfer Learning for Missing Not-at-Random Matrix Completion

Collaborative Compressors in Distributed Mean Estimation with Limited Communication Budge

Distributed Gradient Descent with Many Local Steps in Overparameterized Models