Muchen Li

PhD student at UBC; machine learning and computer vision.

I’m currently a Second Year PHD student at University of British Columbia, Vancouver, supervised by Professor Lenoid Sigal and Professor Renjie Liao. My research interest lies in Machine Learning and Computer Vision.

I received my B.Eng Degree in the Department of Computer Science at Huazhong University of Sci and Tech (HUST) in 2019. After that, I worked as a research assistant at MVIG lab, Shanghai Jiao Tong University (SJTU) from Sep. 2019 to June. 2020 under the supervision of Prof. Cewu Lu (SJTU). I’ve also been worked closely with Dr. Xiangyu Xu (NTU) & Dr. Wenxiu Sun (SenseTime) during my internship at SenseTime.

Publications

  • Jia Jun Cheng Xian*, Muchen Li*, Haotian Yang, Xin Tao, Pengfei Wan, Leonid Sigal, Renjie Liao. Free Lunch Alignment of Text-to-Image Diffusion Models without Preference Image Pairs. In submission. PDF
  • Wenlong Deng, Yi Ren, Muchen Li, Danica J. Sutherland, Xiaoxiao Li, Christos Thrampoulidis. On the Effect of Negative Gradient in Group Relative Deep Reinforcement Optimization. NeurIPS 2025. PDF
  • Qihang Zhang, Muchen Li, Ziao Wang, Renjie Liao, Lele Wang. Neural OOD Text Compression via Test-Time Steering with Weighted Product of Experts. EMNLP 2025. (link forthcoming)
  • Muchen Li*, Sadegh Mahdavi*, Kaiwen Liu, Christos Thrampoulidis, Leonid Sigal, Renjie Liao. AOPS Dataset: Leveraging Online Olympiad-Level Math Problems for LLMs Training and Contamination-Resistant Evaluation. ICML 2025. PDF
  • Muchen Li, Sammy Christen, Chengde Wan, Yujun Cai, Leonid Sigal, Renjie Liao, Shugao Ma. LatentHOI: On the Generalizable Hand Object Motion Generation with Latent Hand Diffusion. CVPR 2025. PDF
  • Xue Yu, Muchen Li, Yan Leng, Renjie Liao. Learning Latent Structures in Network Games via Data-Dependent Gated-Prior Graph Variational Autoencoders. ICML 2024. OpenReview
  • Muchen Li, Jeferrey Liu, Leonid Sigal, Renjie Liao. GraphPNAS: Learning Distribution of Good Neural Architectures via Deep Graph Generative Models. TMLR 2023. OpenReview
  • Muchen Li, Leonid Sigal. Referring Transformer: A One-step Approach to Multi-task Visual Grounding. NeurIPS 2021. Paper
  • Bo Pang, Yizhuo Li, Jiefeng Li, Muchen Li, Hanwen Cao, Cewu Lu. TDAF: Top-Down Attention Framework for Vision Tasks. AAAI 2020. Paper
  • Bo Pang, Yizhuo Li, Muchen Li, Yifan Zhang, Cewu Lu. TubeTK: Adopting Tubes to Track Multi-Object in a Unified One-Step Model. CVPR 2020 (Oral). Paper
  • Xiangyu Xu, Muchen Li, Wenxiu Sun. Learning Spatial and Spatio-Temporal Pixel Aggregations for Image and Video Denoising. IEEE TIP 2020. Paper