Chulhee Yun

13

Papers

6

Total Citations

Papers (13)

Convergence and Implicit Bias of Gradient Descent on Continual Linear Classification

Through the River: Understanding the Benefit of Schedule-Free Methods for Language Model Training

Incremental Gradient Descent with Small Epoch Counts is Surprisingly Slow on Ill-Conditioned Problems

Fundamental Benefit of Alternating Updates in Minimax Optimization

Provable Benefit of Random Permutations over Uniform Sampling in Stochastic Coordinate Descent

Fair Streaming Principal Component Analysis: Statistical and Algorithmic Viewpoint

Practical Sharpness-Aware Minimization Cannot Converge All the Way to Optima

PLASTIC: Improving Input and Label Plasticity for Sample Efficient Reinforcement Learning

Trajectory Alignment: Understanding the Edge of Stability Phenomenon via Bifurcation Theory

Are deep ResNets provably better than linear predictors?

Small ReLU networks are powerful memorizers: a tight analysis of memorization capacity

O(n) Connections are Expressive Enough: Universal Approximability of Sparse Transformers

SGD with shuffling: optimal rates without component convexity and large epoch requirements