CV
Education, research experience, publications, projects, and honors.
Contact Information
| Name | Yuxi Liu |
| Professional Title | Ph.D. Student in Data Science |
| yuxiliu666@stu.pku.edu.cn | |
| Location | Beijing, China |
| Website | https://yuxiliu6666.github.io |
Professional Summary
Ph.D. student at Peking University and research intern at Alibaba Tongyi Wanxiang Lab. I am currently interested in Multimodal Large Language Models, Video Generation, and Video World Models.
Experience
-
2026 - present China
Research Intern
Alibaba Tongyi Wanxiang Lab
Research on multimodal large language models, video generation, and video world models.
Education
-
2023 - 2028 Beijing, China
-
2019 - 2023 Wuhan, China
B.S.
Wuhan University
Mathematics
Interests
Publications
-
2026 CrossDistill: Balancing Quality and Diversity via Trajectory-Level Hybrid Few-Step Distillation
arXiv preprint
Yuxi Liu, Haoyu Li, Yixiang Cai, Tengxu Sun‡, Zekun Zhang, Baole Ai, Ang Wang, Jiamang Wang, Lin Qu, Kun Yuan‡, Kai Zhang‡. * Equal contribution; ‡ Corresponding author. Code
-
2026 RoLA: Rotary-Positioned Low-Rank Linear Attention for Efficient Diffusion Transformers
arXiv preprint
Zekun Zhang, Yixiang Cai, Yuxi Liu*†, Tengxu Sun, Tianle Liu, Zhoutong Wu, Haoyu Li, Baole Ai, Ang Wang, Jiamang Wang, Lin Qu, Kun Yuan‡. * Equal contribution; † Project leader; ‡ Corresponding author. Code
-
2026 RoPeSLR: 3D RoPE-driven Sparse-LowRank Attention for Efficient Diffusion Transformers
arXiv preprint
Yuxi Liu, Zekun Zhang, Yixiang Cai*, Renjia Deng, Yutong He, Kun Yuan‡. * Equal contribution; ‡ Corresponding author.
-
2026 Mixture of Distributions Matters: Dynamic Sparse Attention for Efficient Video Diffusion Transformers
International Conference on Machine Learning (ICML)
Yuxi Liu, Yipeng Hu, Zekun Zhang*, Kunze Jiang, Kun Yuan‡. * Equal contribution; ‡ Corresponding author. Code
-
2026 CR-Net: Scaling Parameter-Efficient Training with Cross-Layer Low-Rank Structure
International Conference on Learning Representations (ICLR)
Boao Kong, Junzhu Liang, Yuxi Liu*, Renjia Deng, Kun Yuan‡. * Equal contribution; ‡ Corresponding author. Code
-
2025 MISA: Memory-Efficient LLMs Optimization with Module-wise Importance Sampling
The Conference on Neural Information Processing Systems (NeurIPS)
Yuxi Liu, Renjia Deng, Yutong He, Xue Wang, Tao Yao, Kun Yuan‡. * Equal contribution; ‡ Corresponding author. Code
Projects
-
CrossDistill
Trajectory-level hybrid few-step distillation that balances generation quality and diversity.
-
RoLA
Rotary-positioned low-rank linear attention for efficient Diffusion Transformers.
-
RoPeSLR
3D RoPE-driven sparse-low-rank attention for efficient Diffusion Transformers.
-
Mixture of Distributions Matters
Dynamic sparse attention for efficient video Diffusion Transformers.
-
CR-Net
Cross-layer low-rank structure for parameter-efficient LLM training.
-
MISA
Module-wise importance sampling for memory-efficient LLM optimization.
Honors and Awards
-
2025 Excellent Research Award
Peking University
-
2023 Outstanding Graduate
Wuhan University
-
2021 National First Prize, Contemporary Undergraduate Mathematical Contest in Modeling
CUMCM
Ranked first in Hubei Province.
-
2021 Meritorious Winner
Mathematical Contest in Modeling / Interdisciplinary Contest in Modeling
-
2021 First Prize, Chinese Mathematics Competitions
Hubei Province
-
2021 Excellent Scholarship for Math Majors (Class I)
Wuhan University
-
2020 First-class Scholarship
Wuhan University
-
2018 Bronze Medal, Chinese Mathematical Olympiad
Chinese Mathematical Society
-
2017 Gold Medal, China Southeast Mathematical Olympiad
China Southeast Mathematical Olympiad