Ziniu Li
 |
Ph.D. student, Email: ziniuli@link.cuhk.edu.cn |
About me
I am a Ph.D. student at The Chinese University of Hong Kong, Shenzhen (CUHKSZ), advised by Prof. Zhi-Quan (Tom) Luo.
I am interested in artificial intelligence, especially reinforcement learning and large language models.
I have worked/interned at Tencent, Nanjing University, Cardinal Operations, etc.
Feel free to contact me if you want to discuss some ideas.
Research Statement
My research focuses on designing adaptive and scalable ML algorithms and analyzing their theoretical guarantees.
In the field of large language models, my work spans several key areas: data selection (NeurIPS 2023, Spotlight), diversity-preserving supervised fine-tuning (NeurIPS 2024 FITML Workshop, Oral), computationally efficient RLHF (ICML 2024), and hallucination mitigation (NeurIPS 2024 AFM Workshop).
In the field of imitation learning, I am interested in the theory of sample complexity (NeurIPS 2020, TPAMI 2021, UAI 2023, Oral), as well as applications in robotics (ICLR 2024 Blog) and signal processing (TSP 2024).
I also work on optimization-centric topics with other researchers, including understanding Adam in training Transformers (NeurIPS 2024), memory-efficient optimizers (ICML 2024 ES-FoMo Workshop), zero-order optimization (IJCAI 2020), and prompt-tuning (EMNLP 2024).
Recent Highlights
*: indicating equal contribution or alphabetic ordering.
Entropic Distribution Matching in Supervised Fine-tuning of LLMs: Less Overfitting and Better Diversity
Ziniu Li, Congliang Chen, Tian Xu, Zeyu Qin, Jiancong Xiao, Ruoyu Sun, Zhi-Quan Luo
Oral Presentation (acceptance rate < 5%), NeurIPS 2024 FITML Workshop
TL;DR: This work introduces a game-theoretic distribution matching method to address the diversity-reducing and knowledge-forgetting issues in SFT
ReMax: A Simple, Effective, and Efficient Reinforcement Learning Method for Aligning Large Language Models
Ziniu Li, Tian Xu, Yushun Zhang, Zhihang Lin, Yang Yu, Ruoyu Sun, Zhi-Quan Luo
The 41st International Conference on Machine Learning (ICML), 2024
TL;DR: This work shows that PPO overshoots for RLHF in LLMs and introduces ReMax, which requires half the memory of PPO and runs twice as fast
When is RL better than DPO in RLHF? A Representation and Optimization Perspective
Ziniu Li* , Tian Xu*, Yang Yu
Oral Presentation, The 12th International Conference on Learning Representations (ICLR) (Tiny Paper Track), 2024
TL;DR: This work analyzes the reward modeling quality in view of representations and the optimization error sources
Imitation Learning from Imperfection: Theoretical Justifications and Algorithms
Ziniu Li* , Tian Xu*, Zeyu Qin, Yang Yu, Zhi-Quan Luo
Spotlight Presentation (acceptance rate < 5%), In Neural Information Processing System (NeurIPS) 37, 2023
TL;DR: This work validates that importance sampling is effective in data selection when leveraging multiple imperfect (out-of-distribution and low-quality) data sources
Service
Reviewer
NeurIPS (Top Reviewer), ICML (Outstanding Reviewer), ICLR (Highlighted Reviewer).
Teaching Assistant
DDA6111: Discrete Optimization. 2022 Spring @ CUHKSZ
DDA6060: Machine Learning. 2023 Spring @ CUHKSZ
FTE4560: Basic Machine Learning. 2021 Spring @ CUHKSZ.
CSC4120: Design and Analysis of Algorithms. 2022 Fall, 2021 Fall @ CUHKSZ
MAT3007: Introduction to Optimization. 2020 Fall @ CUHKSZ
Lecturer
Machine Learning (Summer Course for Senior High School Students) @ X ACADEMY 2022 TechX
Award
[2024-01] Runner-up of poster presentation award at the third doctoral and postdoctoral forum of Shenzhen Research Institute of Big Data. $5,000 CNY
[2023-12] Guo-Tai-Jun-An Scholarship. $20,000 CNY
[2021-04] Best oral presentation award at the first doctoral and postdoctoral forum of Shenzhen Research Institute of Big Data. $5,000 CNY