One model for molecules, crystals and proteins.
I'm Yuxuan Ren, a Ph.D. student at the National University of Singapore. I build generative models and molecular languages for science.
About
I work on AI for Science at the School of Computing, National University of Singapore, advised by Yatao Bian in the Blue Whale Lab.
My recent work asks how far one plain Transformer can go across atomic systems. SinAE is a single autoencoder that reconstructs molecules, crystals and proteins almost losslessly. STAMP is a molecular language that makes fragments easy to learn from and to edit.
I trained as a chemist at USTC before moving into machine learning, with time at Microsoft Research AI for Science and the Beijing Academy of Artificial Intelligence along the way.
-
2026 to now
Ph.D. student
National University of Singapore, School of Computing
Advised by Yatao Bian, Blue Whale Lab -
2026
ZGCM-1 post-training
ZGCM Team
Mentored by Jiyan He -
2024 to 2025
Researcher
Beijing Academy of Artificial Intelligence
Supervised by Qiwei Ye -
2023 to 2024
Research intern
Microsoft Research AI for Science
Supervised by Chang Liu -
2024
B.S. in Chemistry
University of Science and Technology of China
Advised by Zhonghuai Hou
Selected papers
All publications-
2026
STAMP as a Topology-Aware Molecular Language for Fragment-Level Learning and Editing
-
2026
ZGCM-1: A Fully Open and Extremely Efficient Foundation Model for Math and Agentic Search
-
2026
SinAE: A Single-Architecture Flow-Matching Autoencoder for Cross-Domain Atomic Systems
-
2026
EPO: Diverse and Realistic Protein Ensemble Generation via Energy Preference Optimization
-
2025
CryoGEN: Energy-based Cryo-ET Reconstruction
-
2024
Physical Consistency Bridges Heterogeneous Data in Molecular Multi-Task Learning
† Equal contribution.
News
- Oct 2026STAMP, a topology-aware molecular language, is on arXiv.
- Sep 2026ZGCM-1, a fully open 7B foundation model for math and agentic search, is released.
- Jul 2026SinAE and Chemical Chain-of-Thought are on arXiv.
- Nov 2025EPO is accepted at AAAI 2026.
- Mar 2025CryoGEN is accepted at ICLR 2025.
- Sep 2024Physical Consistency MTL and Neural P3M are accepted at NeurIPS 2024.