Yuxuan Song「宋宇轩」

I am a last-year Ph.D. student in the Department of Computer Science and Technology at Tsinghua University, affiliated with the Institute for Artificial Intelligence Industry Research (AIR), advised by Prof. Wei-Ying Ma. Previously, I used to be a researcher at Bytedance AI Lab, where I worked on deep generative models as well as their application in structured data with Prof. Hao Zhou and Prof. Lei Li. Before that, I received my bachelor and master degree from Shanghai Jiao Tong University (SJTU) where I am fortunate to be advised by Prof. Yong Yu. I am currently on the industry job market.

Email  /  CV  /  Google Scholar  /  Twitter  /  Github

My ultimate goal is to integrate generalizable intelligence (LLMs) with advanced AI tools (e.g., AF3) to solve challenging real-world scientific problems and boost scientific discovery (LLM4Science). My Ph.D. research has focused on geometric generative models (bio/chemical/material structures), discrete diffusion models (biological sequences, Diffusion LLMs), and reasoning in LLMs.

Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inference[pdf][demo]
Yuxuan Song*, Zheng Zhang*, Cheng Luo*, Pengyang Gao, Fan Xia, Hao Luo, Zheng Li, Yuehang Yang, Hongli Yu, Xingwei Qu, Yuwei Fu, Jing Su, Ge Zhang, Wenhao Huang, Mingxuan Wang, Lin Yan, Xiaoying Jia, Jingjing Liu, Wei-Ying Ma, Ya-Qin Zhang, Yonghui Wu, Hao Zhou
Technical Report, 2025.

ShortListing Model: A Streamlined SimplexDiffusion for Discrete Variable Generation[pdf]
Yuxuan Song*, Zhe Zhang*, Yu Pei*, Jingjing Gong, Qiying Yu, Zheng Zhang, Mingxuan Wang, Hao Zhou, Jingjing Liu, Wei-Ying Ma
Neural Information Processing Systems (NeurIPS), 2025.

Accelerating 3D Molecule Generative Models with Trajectory Diagnosis[pdf]
Zhilong Zhang*, Yuxuan Song*, Yichun Wang*, Jingjing Gong, Hanlin Wu, Dongzhan Zhou, Hao Zhou, Wei-Ying Ma
Neural Information Processing Systems (NeurIPS), 2025.

Smooth Interpolation for Improved Discrete Graph Generative Models[pdf]
Yuxuan Song*, Juntong Shi*, Jingjing Gong*, Minkai Xu, Stefano Ermon, Hao Zhou, Wei-Ying Ma
International Conference on Machine Learning (ICML), 2025.

Piloting Structure-Based Drug Design via Modality-Specific Optimal Schedule[pdf]
Keyue Qiu*, Yuxuan Song*, Zhehuan Fan, Peidong Liu, Zhe Zhang, Mingyue Zheng, Hao Zhou, Wei-Ying Ma
International Conference on Machine Learning (ICML), 2025.

Steering Protein Family Design through Profile Bayesian Flow [pdf]
Jingjing Gong*, Yu Pei*, Siyu Long*, Yuxuan Song*, Zhe Zhang, Wenhao Huang, Ziyao Cao, Shuyi Zhang, Hao Zhou, Wei-Ying Ma
International Conference on Learning Representations (ICLR), 2025.
Oral Presentation [207/11672]

A Periodic Bayesian Flow for Material Generation [pdf][code]
Hanlin Wu*, Yuxuan Song*, Jingjing Gong, Ziyao Cao, Yawen Ouyang, Jianbing Zhang, Hao Zhou, Wei-Ying Ma, Jingjing Liu
International Conference on Learning Representations (ICLR), 2025.
Spotlight Paper

RetroDiff: Retrosynthesis as Multi-stage Distribution Interpolation[pdf]
Yiming Wang, Yuxuan Song, Minkai Xu, Rui Wang, Hao Zhou, Weiying Ma
International Conference on Artificial Intelligence and Statistics (AISTATS), 2025.

MolCRAFT: Structure-Based Drug Design in Continuous Parameter Space [pdf][code]
Yanru Qu*, Keyue Qiu*, Yuxuan Song*, Jingjing Gong, Jiawei Han, Mingyue Zheng, Hao Zhou, Weiying Ma
International Conference on Machine Learning (ICML), 2024.

Unified Generative Modeling of 3D Molecules with Bayesian Flow Networks [pdf][code]
Yuxuan Song*, Jingjing Gong*, Yanru Qu, Hao Zhou, Mingyue Zheng, Jingjing Liu, Weiying Ma
International Conference on Learning Representations (ICLR), 2024.
Oral Presentation [85/7262]

Equivariant Flow Matching with Hybrid Probability Transport for 3D Molecule Generation [pdf][code]
Yuxuan Song*, Jingjing Gong*, Ziyao Cao, Minkai Xu, Stefano Ermon, Hao Zhou, Weiying Ma
Neural Information Processing Systems (NeurIPS), 2023.

Coarse-to-Fine: a Hierarchical Diffusion Model for Molecule Generation in 3D [pdf][code]
Bo Qiang*, Yuxuan Song*, Minkai Xu, Bowen Gao, Hao Zhou, Weiying Ma, Yanyan Lan
International Conference on Machine Learning (ICML), 2023.

Follow Your Path: a Progressive Method for Knowledge Distillation. [pdf]
Wenxian Shi*, Yuxuan Song*,Hao Zhou, Lei Li
European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD), 2021

Improving Maximum Likelihood Training for Text Generation with Density Ratio Estimation. [pdf]
Yuxuan Song, Ning Miao, Hao Zhou, Lantao Yu, Mingxuan Wang, Lei Li
Artificial Intelligence and Statistics (AISTATS) 2020.

Infomax Neural Joint Source-Channel Coding via Adversarial Bit Flip. [pdf][code]
Yuxuan Song, Minkai Xu, Lantao Yu, Hao Zhou, Shuo Shao, Yong Yu
AAAI Conference on Artificial Intelligence (AAAI), 2020.

I am a die-hard fan of Liaoning Flying Leopard. I adopted a lovely Ragdoll (many thanks to Changzhi Sun).