GuangChen Li

GuangChen Li （李光宸）

Portrait
M.S. Student in Data Science
University of Michigan, Ann Arbor

I am an M.S. student in Data Science at the University of Michigan, Ann Arbor, with expected graduation in May 2027. My research focuses on generative modeling, time series, and world models.

I received my B.S. in Physics from Renmin University of China in 2025 and was a visiting student at the University of California, Davis in 2024. Feel free to contact me if you are interested in related research or potential collaborations.

GitHub LinkedIn

Education

  • University of Michigan, Ann Arbor

  • Renmin University of China

  • University of California, Davis

Awards

  • Merit Student 2024
  • Second Prize, National College Student Mathematics Competition 2024
  • Academic Progress Scholarship 2023

News

2026
Our paper Raster or Random? Let the Data Decide was submitted to ICLR 2027.
Sep 2026
Our work on generative adaptation of temporal point processes was submitted to AAAI 2027.
Jul 2026
I joined Prof. Qing Qu's DeepThink Lab at the University of Michigan to work on diffusion models.
Apr 2026
2025
I began collaborating remotely with Prof. Hongteng Xu at the Gaoling School of Artificial Intelligence, Renmin University of China, on generative adaptation for temporal point processes.
Oct 2025
I joined Prof. Raed Al Kontar's group as a research assistant, working on reliable LLM agent systems.
Sep 2025
I started the M.S. in Data Science program at the University of Michigan, Ann Arbor.
Aug 2025
I graduated from Renmin University of China with a bachelor's degree. 🎓
Jun 2025

Publications (view all )

Raster or Random? Let the Data Decide
Raster or Random? Let the Data Decide

Lianghe Shi*, GuangChen Li*, Huijie Zhang, Yixuan Jia, Qing Qu (* equal contribution)

Under review

The best generation order depends on the data’s dependency structure: autoregressive order performs better on language and ordinary LDS, while random order performs better on images and shuffled LDS. A unified framework isolates order’s effects and explains its distinct roles during training and inference.

Raster or Random? Let the Data Decide
Raster or Random? Let the Data Decide

Lianghe Shi*, GuangChen Li*, Huijie Zhang, Yixuan Jia, Qing Qu (* equal contribution)

Under review

The best generation order depends on the data’s dependency structure: autoregressive order performs better on language and ordinary LDS, while random order performs better on images and shuffled LDS. A unified framework isolates order’s effects and explains its distinct roles during training and inference.

Generative Adaptation of Temporal Point Processes for Generalizable Event Sequence Prediction and Generation
Generative Adaptation of Temporal Point Processes for Generalizable Event Sequence Prediction and Generation

Qingmei Wang*, GuangChen Li*, Yitian Wang, Tianyu Huang, Qianya Wang, Xiaobo Qin, Shulin Wang, YangHua, Hongteng Xu (* equal contribution)

Under review

GA-TPP generates condition-specific LoRA adapters with conditional diffusion, enabling pretrained temporal point processes to adapt to unseen sequence contexts and distribution shifts.

Generative Adaptation of Temporal Point Processes for Generalizable Event Sequence Prediction and Generation
Generative Adaptation of Temporal Point Processes for Generalizable Event Sequence Prediction and Generation

Qingmei Wang*, GuangChen Li*, Yitian Wang, Tianyu Huang, Qianya Wang, Xiaobo Qin, Shulin Wang, YangHua, Hongteng Xu (* equal contribution)

Under review

GA-TPP generates condition-specific LoRA adapters with conditional diffusion, enabling pretrained temporal point processes to adapt to unseen sequence contexts and distribution shifts.

All publications