I am a CS Ph.D. student at UNC Chapel Hill, supervised by Prof. Mingyu Ding and Prof. Huaxiu Yao. I received both my Bachelorโ€™s and Masterโ€™s degrees from Tongji University. I interned at UC Berkeley as a visiting student researcher, where I collaborated with Prof. Masayoshi Tomizuka, Prof. Mingyu Ding, and Dr. Wei Zhan.

My research interests include vision-language model, RL for reasoning models , and embodied AI. And my research goal is to improve the performance of LLMs and VLMs through reinforcement learning and other post-training techniques, and to leverage these models to build efficient and intelligent agents capable of interacting with the physical world (embodied intelligence) and driving scientific exploration (AI for Science).

I am actively seeking 2026 summer internships. Letโ€™s connect!

๐Ÿ”ฅ News

  • 2025.11: ย ๐ŸŽ‰๐ŸŽ‰ One paper is accepted by AAAI 2026.
  • 2025.10: ย  Alignment Tipping Process: How Self-Evolution Pushes LLM Agents Off the Rails is preprinted.
  • 2025.09: ย  We released SciReasoner, a scientific reasoning foundation model covering 103 tasks. It is pretrained on a 206B-token corpus and further enhanced through supervised fine-tuning and reinforcement learning to continually improve its reasoning ability. The model achieves state-of-the-art performance on more than 50 tasks. Paper, Github, Hugging Face
  • 2025.08: ย  We released A Survey of Scientific Large Language Models: From Data Foundations to Agent Frontiers, providing a comprehensive, data-centric roadmap of Sci-LLMs from foundational models to agentic frontiers. Paper, Github
  • 2025.08: ย  We released VIPER-R1, a multimodal model for Visual Induction for Physics-based Equation Reasoning to discover fundamental symbolic formulas. Paper, Project.
  • 2025.06: ย  PhysUniBench is released, which is a large-scale multimodal benchmark for evaluating undergraduate-level physics reasoning in AI models. Paper, Project.
  • 2025.06: ย ๐ŸŽ‰๐ŸŽ‰ I received my Master degree from Tongji University, and got the honor of Shanghai Outstanding Graduate(Top 1%)!
  • 2022.09: ย ๐ŸŽ‰๐ŸŽ‰ I was bestowed with the Nominee Award for Shanghai University Student Annual Character, a recognition of paramount importance in evaluating the comprehensive aptitudes and societal influence of college students within the Shanghai municipality. Only 20 people in Shanghai receive the award each year (including nominees)!
  • 2021.05: ย ๐ŸŽ‰๐ŸŽ‰ I received the Tongji University Pursuit of Excellence Award, the highest honor for undergraduates at Tongji University! Only three undergraduates are able to receive this honor each year!
  • 2020.10: ย ๐ŸŽ‰๐ŸŽ‰ I won the Best Paper Award in CUMCM! CUMCM is the largest basic academic competition for undergraduate students in China, and 3 papers were awarded from 42,000+ candidates!

๐Ÿ“ Publications

sym

Mimicking the Physicistโ€™s Eye:A VLM-centric Approach for Physics Formula Discovery (VIPER-R1)

Jiaqi Liu, Songning Lai, Pengze Li, Di Yu, Wenjie Zhou, Yiyang Zhou, Peng Xia, Zijun Wang, Xi Chen, Shixiang Tang, Lei Bai, Wanli Ouyang, Mingyu Ding, Huaxiu Yao, Aoran Wang, accepted by NeurIPS 2025 Efficient Reasoning Workshop.

Sci-LLM Survey

SciReasoner: Laying the Scientific Reasoning Ground Across Disciplines

Yizhou Wang$^\star$, Chen Tang$^\star$, Han Deng$^\star$, Jiabei Xiao$^\star$, Jiaqi Liu$^\star$, Jianyu Wu$^\star$, โ€ฆ, Philip Torr, Shixiang Tang, Xinzhu Ma, Wanli Ouyang, Lei Bai et al. . arXiv:2509.21320.

sym

Alignment Tipping Process: How Self-Evolution Pushes LLM Agents Off the Rails

Siwei Han, Jiaqi Liu, Yaofeng Su, Wenbo Duan, Xinyuan Liu, Cihang Xie, Mohit Bansal, Mingyu Ding, Linjun Zhang, Huaxiu Yao, arxiv:2510.04860.

Sci-LLM Survey

A Survey of Scientific Large Language Models: From Data Foundations to Agent Frontiers

Ming Hu, Chenglong Ma, Wei Li, Wanghan Xu, Jiamin Wu, Jucheng Hu, Tianbin Li, Guohang Zhuang, Jiaqi Liu, Yingzhou Lu, โ€ฆ, Wanli Ouyang, Yu Qiao, Zongyuan Ge, Shixiang Tang, Junjun He et al. . arxiv:2508.21148.

sym

Language-Driven Policy Distillation for Cooperative Driving in Multi-Agent Reinforcement Learning

Jiaqi Liu, Chengkai Xu, Peng Hang, Jian Sun, Mingyu Ding, Wei Zhan, Masayoshi Tomizuka, IEEE Robotics and Automation Letters (RA-L), DOI: 10.1109/LRA.2025.3551098.

sym

DDM-Lag : A Diffusion-based Decision-making Model for Autonomous Vehicles with Lagrangian Safety Enhancement

Jiaqi Liu, Peng Hang, Xiaocong Zhao, Jianqiang Wang, Jian Sun, IEEE Transactions on Artificial Intelligence(TAI), DOI: 10.1109/TAI.2024.3497918.

sym

Cooperative Decision-Making for CAVs at Unsignalized Intersections: A MARL Approach with Attention and Hierarchical Game Priors

Jiaqi Liu, Peng Hang, Xiaoxiang Na, Chao Huang, Jian Sun, IEEE Transactions on Intelligent Transportation Systems (TITS), DOI: 10.1109/TITS.2024.3503092, 2024.

๐ŸŽ– Honors and Awards

๐Ÿ“– Educations

  • 2025.08 - now, PhD, UNC Chapel Hill, NC, US
  • 2022.09 - 2025.06, Master, Tongji University, Shanghai, China
  • 2018.09 - 2022.06, Bachelor, Tongji University, Shanghai, China (GPA: 4.89/5, rank: 1/163)

๐Ÿ“š Academic Services

Reviews

  • Conference Reviewer: NeurIPS, ICML, ICCV, AAAI, ICRA, ITSC
  • Journal Reviewer: IEEE Transactions on Intelligent Vehicles (TIV), IEEE Transactions on Intelligent Transportation Systems (TITS), IEEE Transactions on Neural Networks and Learning(TNNLS),IEEE Robotics and Automation Letters (RA-L), Journal of Field Robotics, IEEE Transactions on Industrial Informatics (TII), IEEE Transactions on Vehicular Technology (TVT), IEEE Transactions on Automation Science and Engineering (TASE), IEEE Internet of Things Journal, Nonlinear Dynamics, Journal of Advanced Transportation, Scientific Reports

Mentoring

  • Yaofeng Su: Fudan University
  • Kaiwen Xiong: Shanghai Jiao Tong University
  • Carsen Sharkey: UNC Chapel Hill
  • Yicheng Guo: Tongji University
  • Chengkai Xu: Tongji University
  • Yuhang Zhang: Tongji University

Teaching Assistant

  • CS 790-183: Transfer Learning, UNC, Fall 2025