Paper 'Elucidating Rectified Flow with Deterministic Sampler' accepted by CSML 2025
Paper 'Improved Discretization Complexity Analysis of Consistency Models' accepted by ICML 2025
Paper 'The Polynomial Iteration Complexity for Variance Exploding Diffusion Models' accepted by AISTATS 2025
Two papers on iteration complexity and few-shot diffusion models accepted by NeurIPS 2024; one received 2nd Prize Best Paper Award at TongAI 2025
Paper 'Understanding Representation Learnability of Nonlinear Self-Supervised Learning' accepted by AAAI 2023 (Oral)
Paper 'Learning Adversarial Linear Mixture Markov Decision Processes...' accepted by ICLR 2023
Paper 'Learning Adversarial Low-rank Markov Decision Processes...' accepted by NeurIPS 2023
Multiple papers submitted to ICML 2025, SODA 2026, etc.
Serving as reviewer for ICML 2025 and AISTATS 2025
Background
Third-year PhD student in Computer Science at Shanghai Jiao Tong University
Affiliated with the John Hopcroft Center, supervised by Prof. Shuai Li since September 2022
Research interests: diffusion models, deep learning theory, reinforcement learning, and RLHF
Currently focusing on theoretical analysis of diffusion models (sampling complexity, statistical complexity, optimization perspective)
Exploring the integration of diffusion models and reinforcement learning, e.g., RL-based fine-tuning of diffusion models and using diffusion in offline-to-online RL