Jinhan Li

photo.jpg

Yao Class, IIIS, Tsinghua

lijinhan21@mails.tsinghua.edu.cn

Photo taken by Zhuorui Ye.

I am Jinhan Li (李谨菡), a first-year PhD student at IIIS, Tsinghua University, advised by Professor Kaifeng Lyu. I spent my undergraduate years in Yao Class, directed by Andrew Chi-Chih Yao. I was very fortunate to work as a student researcher in Professor Yuke Zhu’s group at UT Austin and Professor Yi Wu’s group at IIIS.

My research interests lie broadly in machine learning. I’m attracted to research that integrates empirical observations with rigorous experiments and theoretical analysis, offering insights into modern machine learning techniques. Currently, I’m exploring data-centric methods in LLM pretraining, as well as trying to understand the training dynamics of RLVR. Previously, I worked on robot learning.

I’m open to chatting about research or life. Feel free to drop me an email if you are interested!

Selected Publications

2026

  1. srp.png
    Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection
    Jinhan Li*, Kexian Tang*, Yihan Xu, Zhuorui Ye, and Kaifeng Lyu
    In Trustworthy AI for Good (AI4GOOD) Workshop @ ICML 2026, 2026
    arXiv

2024

  1. okami.png
    OKAMI: Teaching Humanoid Robots Manipulation Skills through Single Video Imitation
    Jinhan Li, Yifeng Zhu*, Yuqi Xie*, Zhenyu Jiang*, Mingyo Seo, Georgios Pavlakos, and Yuke Zhu
    In 8th Annual Conference on Robot Learning (CoRL), 2024

    Oral Presentation

    arXiv Website
  2. musical.png
    Sing it, Narrate it: Quality Musical Lyrics Translation
    Zhuorui Ye*, Jinhan Li*, and Rongwu Xu
    In Findings of the Association for Computational Linguistics: EMNLP 2024, 2024
    arXiv Website
  3. li2023learning.png
    Learning Agile Bipedal Motions on a Quadrupedal Robot
    Yunfei Li, Jinhan Li, Wei Fu, and Yi Wu
    In 2024 IEEE International Conference on Robotics and Automation (ICRA), 2024

    ICRA EXPO Best Demo Finalist

    arXiv
  4. ouyang2024longhorizon.png
    Long-horizon Locomotion and Manipulation on a Quadrupedal Robot with Large Language Models
    Yutao Ouyang*, Jinhan Li*, Yunfei Li, Zhongyu Li, Chao Yu, Koushil Sreenath, and Yi Wu
    2024
    arXiv Website