Yuejie Chi

24

Papers

65

Total Citations

Papers (24)

ShadowKV: KV Cache in Shadows for High-Throughput Long-Context LLM Inference

Breaking the Curse of Multiagency in Robust Multi-Agent Reinforcement Learning

A Theoretical Analysis of Self-Supervised Learning for Vision Transformers

Transformers Provably Learn Chain-of-Thought Reasoning with Length Generalization

Understanding Masked Autoencoders via Hierarchical Latent Variable Models

Get More with LESS: Synthesizing Recurrence with KV Cache Compression for Efficient LLM Inference

Accelerating Convergence of Score-Based Diffusion Models, Provably

Sample-Efficient Robust Multi-Agent Reinforcement Learning in the Face of Environmental Uncertainty

Federated Offline Reinforcement Learning: Collaborative Single-Policy Coverage Suffices

Identification of Nonlinear Latent Hierarchical Models

Reward-agnostic Fine-tuning: Provable Statistical Benefits of Hybrid Reinforcement Learning

Counterfactual Generation with Identifiability Guarantees

Seeing is not Believing: Robust Reinforcement Learning against Spurious Correlation

The Curious Price of Distributional Robustness in Reinforcement Learning with a Generative Model

Provable Non-convex Phase Retrieval with Outliers: Median TruncatedWirtinger Flow

Implicit Regularization in Nonconvex Statistical Estimation: Gradient Descent Converges Linearly for Phase Retrieval and Matrix Completion

Sample Complexity of Asynchronous Q-Learning: Sharper Analysis and Variance Reduction

Breaking the Sample Size Barrier in Model-Based Reinforcement Learning with a Generative Model

Sample-Efficient Reinforcement Learning Is Feasible for Linearly Realizable MDPs with Limited Revisiting

Breaking the Sample Complexity Barrier to Regret-Optimal Model-Free Reinforcement Learning

Fast Policy Extragradient Methods for Competitive Games with Entropy Regularization

SoteriaFL: A Unified Framework for Private Federated Learning with Communication Compression

Minimax-Optimal Multi-Agent RL in Markov Games With a Generative Model

BEER: Fast $O(1/T)$ Rate for Decentralized Nonconvex Optimization with Communication Compression