Michael Crawshaw

21 papers A* 9Journal 12

Year	Rank	Type	Title / Venue / Authors
2026	J	`jnl`	Non-Euclidean Gradient Descent Operates at the Edge of Stability. CoRR Rustem Islamov, Michael Crawshaw, Jeremy Cohen, Robert M. Gower
2026	J	`jnl`	Tight Bounds for Logistic Regression with Large Stepsize Gradient Descent in Low Dimension. CoRR Michael Crawshaw, Mingrui Liu
2025	J	`jnl`	An Exploration of Non-Euclidean Gradient Descent: Muon and its Many Variants. CoRR Michael Crawshaw, Chirag Modi, Mingrui Liu, Robert M. Gower
2025	A*	`conf`	Complexity Lower Bounds of Adaptive Gradient Algorithms for Non-convex Stochastic Optimization under Relaxed Smoothness. ICLR Michael Crawshaw, Mingrui Liu
2025	J	`jnl`	Complexity Lower Bounds of Adaptive Gradient Algorithms for Non-convex Stochastic Optimization under Relaxed Smoothness. CoRR Michael Crawshaw, Mingrui Liu
2025	A*	`conf`	Constant Stepsize Local GD for Logistic Regression: Acceleration by Instability. ICML Michael Crawshaw, Blake Woodworth, Mingrui Liu
2025	J	`jnl`	Constant Stepsize Local GD for Logistic Regression: Acceleration by Instability. CoRR Michael Crawshaw, Blake Woodworth, Mingrui Liu
2025	A*	`conf`	Local Steps Speed Up Local GD for Heterogeneous Distributed Logistic Regression. ICLR Michael Crawshaw, Blake Woodworth, Mingrui Liu
2025	J	`jnl`	Local Steps Speed Up Local GD for Heterogeneous Distributed Logistic Regression. CoRR Michael Crawshaw, Blake Woodworth, Mingrui Liu
2024	A*	`conf`	Federated Learning under Periodic Client Participation and Heterogeneous Data: A New Communication-Efficient Algorithm and Analysis. NeurIPS Michael Crawshaw, Mingrui Liu
2024	J	`jnl`	Federated Learning under Periodic Client Participation and Heterogeneous Data: A New Communication-Efficient Algorithm and Analysis. CoRR Michael Crawshaw, Mingrui Liu
2024	A*	`conf`	Provable Benefits of Local Steps in Heterogeneous Federated Learning for Neural Networks: A Feature Learning Perspective. ICML Yajie Bao, Michael Crawshaw, Mingrui Liu
2023	A*	`conf`	EPISODE: Episodic Gradient Clipping with Periodic Resampled Corrections for Federated Learning with Heterogeneous Data. ICLR Michael Crawshaw, Yajie Bao, Mingrui Liu
2023	J	`jnl`	EPISODE: Episodic Gradient Clipping with Periodic Resampled Corrections for Federated Learning with Heterogeneous Data. CoRR Michael Crawshaw, Yajie Bao, Mingrui Liu
2023	A*	`conf`	Federated Learning with Client Subsampling, Data Heterogeneity, and Unbounded Smoothness: A New Algorithm and Lower Bounds. NeurIPS Michael Crawshaw, Yajie Bao, Mingrui Liu
2022	A*	`conf`	Fast Composite Optimization and Statistical Recovery in Federated Learning. ICML Yajie Bao, Michael Crawshaw, Shan Luo, Mingrui Liu
2022	J	`jnl`	Fast Composite Optimization and Statistical Recovery in Federated Learning. CoRR Yajie Bao, Michael Crawshaw, Shan Luo, Mingrui Liu
2022	A*	`conf`	Robustness to Unbounded Smoothness of Generalized SignSGD. NeurIPS Michael Crawshaw, Mingrui Liu, Francesco Orabona, Wei Zhang, Zhenxun Zhuang
2022	J	`jnl`	Robustness to Unbounded Smoothness of Generalized SignSGD. CoRR Michael Crawshaw, Mingrui Liu, Francesco Orabona, Wei Zhang, Zhenxun Zhuang
2021	J	`jnl`	SLAW: Scaled Loss Approximate Weighting for Efficient Multi-Task Learning. CoRR Michael Crawshaw, Jana Kosecká
2020	J	`jnl`	Multi-Task Learning with Deep Neural Networks: A Survey. CoRR Michael Crawshaw