Xueyi Chen

悟已往之不谏

知来者之可追

Hangzhou, Zhejiang, China

About Me

I am Xueyi Chen, a first-year PhD student at Westlake University, advised by Prof. Huan Wang at the ENCODE Lab. I am currently a research intern at inclusionAI, Ant Group. Previously, I earned my MSc in Computer Science from The Chinese University of Hong Kong (CUHK) in 2025, supervised by Prof. Bei Yu, and my BEng in Software Engineering from Fuzhou University in 2024.

Research Interests

I am broadly interested in building efficient and capable AI systems:

Multimodal Understanding [StreamingTOM] [LVOmniBench]
LLMs for Kernel Generation [MetaKernelBench]
RL for LLM Reasoning [OORL] [ToTRL]

I am always open to potential collaborations. Please feel free to contact me at xueyi.chen.2024@gmail.com or chenxueyi@westlake.edu.cn.

news

Oct 06, 2026	📄 Our preprint MetaKernelBench on LLM-driven kernel generation is now on arXiv! Check out the paper, project page, and code. Thanks to all co-authors!
Aug 21, 2026	🎓 Started my PhD at Westlake University, advised by Prof. Huan Wang at the ENCODE Lab. Excited for the journey ahead!
Mar 19, 2026	📄 We released LVOmniBench, a comprehensive benchmark for evaluating omnimodal LLMs on long audio-video understanding! This work was led by co-first authors Keda Tao and Yuhua Zheng. Check out our paper, project page, and code!
Feb 21, 2026	🎉 StreamingTOM has been accepted to CVPR 2026! This is my first top-venue paper as first author! Thanks to all my co-authors, and see you in Denver! ✝ Special thanks to PTE.* for advice on figure design and for encouragement during that time!*
Oct 21, 2025	📄 Our paper “StreamingTOM: Streaming Token Compression for Efficient Video Understanding” is now on arXiv! Thanks to my co-authors Keda Tao, Kele Shao, and Prof. Huan Wang. Check it out here!

selected publications

arXiv

MetaKernelBench: Measuring GPU Kernel Knowledge Transfer Beyond Code

Xueyi Chen^*, Shiyu Liu^*, Xin Jin, Yuhua Zheng, Xin Li, Haolei Bai, Junhan Zhu, and Huan Wang

arXiv preprint arXiv:2610.05014, Oct 2026

Under review

arXiv Bib PDF Code Website

@article{chen2026metakernelbench,
  title = {MetaKernelBench: Measuring GPU Kernel Knowledge Transfer Beyond Code},
  author = {Chen, Xueyi and Liu, Shiyu and Jin, Xin and Zheng, Yuhua and Li, Xin and Bai, Haolei and Zhu, Junhan and Wang, Huan},
  journal = {arXiv preprint arXiv:2610.05014},
  year = {2026},
  month = oct,
  note = {Under review},
}

CVPR

StreamingTOM: Streaming Token Compression for Efficient Video Understanding

Xueyi Chen, Keda Tao, Kele Shao, and Huan Wang

In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun 2026

arXiv Bib PDF Code Website

@inproceedings{chen2026streamingtom,
  title = {StreamingTOM: Streaming Token Compression for Efficient Video Understanding},
  author = {Chen, Xueyi and Tao, Keda and Shao, Kele and Wang, Huan},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2026},
  month = jun,
}

arXiv

ToTRL: Unlock LLM Tree-of-Thoughts Reasoning Potential through Puzzles Solving

Haoyuan Wu^*, Xueyi Chen^*, Rui Ming, Jilong Gao, Shoubo Hu, Zhuolun He, and Bei Yu

arXiv preprint arXiv:2505.12717, May 2025

Under review

arXiv Bib PDF

@article{wu2025totrl,
  title = {ToTRL: Unlock LLM Tree-of-Thoughts Reasoning Potential through Puzzles Solving},
  author = {Wu, Haoyuan and Chen, Xueyi and Ming, Rui and Gao, Jilong and Hu, Shoubo and He, Zhuolun and Yu, Bei},
  journal = {arXiv preprint arXiv:2505.12717},
  year = {2025},
  month = may,
  note = {Under review},
}