Picture for Quanquan Gu

Quanquan Gu

Variance-Dependent Regret Lower Bounds for Contextual Bandits

Add code
Mar 15, 2025
Viaarxiv icon

Global Convergence and Rich Feature Learning in $L$-Layer Infinite-Width Neural Networks under $μ$P Parametrization

Add code
Mar 12, 2025
Viaarxiv icon

Energy-Weighted Flow Matching for Offline Reinforcement Learning

Add code
Mar 06, 2025
Viaarxiv icon

Understanding SGD with Exponential Moving Average: A Case Study in Linear Regression

Add code
Feb 19, 2025
Viaarxiv icon

Multi-Step Alignment as Markov Games: An Optimistic Online Gradient Descent Approach with Convergence Guarantees

Add code
Feb 18, 2025
Viaarxiv icon

Logarithmic Regret for Online KL-Regularized Reinforcement Learning

Add code
Feb 11, 2025
Viaarxiv icon

Nearly Optimal Sample Complexity of Offline KL-Regularized Contextual Bandits under Single-Policy Concentrability

Add code
Feb 09, 2025
Viaarxiv icon

Tensor Product Attention Is All You Need

Add code
Jan 11, 2025
Viaarxiv icon

Towards Simple and Provable Parameter-Free Adaptive Gradient Methods

Add code
Dec 27, 2024
Figure 1 for Towards Simple and Provable Parameter-Free Adaptive Gradient Methods
Figure 2 for Towards Simple and Provable Parameter-Free Adaptive Gradient Methods
Figure 3 for Towards Simple and Provable Parameter-Free Adaptive Gradient Methods
Figure 4 for Towards Simple and Provable Parameter-Free Adaptive Gradient Methods
Viaarxiv icon

MARS: Unleashing the Power of Variance Reduction for Training Large Models

Add code
Nov 15, 2024
Figure 1 for MARS: Unleashing the Power of Variance Reduction for Training Large Models
Figure 2 for MARS: Unleashing the Power of Variance Reduction for Training Large Models
Figure 3 for MARS: Unleashing the Power of Variance Reduction for Training Large Models
Figure 4 for MARS: Unleashing the Power of Variance Reduction for Training Large Models
Viaarxiv icon