I am Xiaoyang Hou (侯晓阳 in Chinese), a PhD candidate at École Polytechnique Fédérale de Lausanne (EPFL), co-supervised by Patrick Barth and Pierre Vandergheynst. I explore deep learning for biomolecular design and discovery.
My research focuses on unified biomolecule design methods and foundation models. I develop machine learning methods at the intersection of biology and chemistry, with a particular interest in computational drug design and discovery.
I welcome opportunities for collaboration and academic discussion. Please feel free to reach out if you are interested in my research or have questions to explore.
Publications
View all publications
TMLR 2025 · NeurIPS AIDrugX Workshop 2024
GGFlow: A Graph Flow Matching Method with Efficient Optimal Transport
Discrete flow matching with efficient optimal transport for graph generation. An edge-augmented graph transformer and a goal-guided framework steer generation toward desired properties.
Bioinformatics 2024
GTAM: A Molecular Pretraining Model with Geometric Triangle Awareness
Geometry-aware molecular encoders connect 2D topology and 3D conformations. Two contrastive objectives transfer edge information between these views to improve molecular representations.
* Equal contribution.
Education
Oct 2026 — Present
École Polytechnique Fédérale de Lausanne (EPFL)
Laboratory of Protein and Cell Engineering
PhD Candidate
Sep 2022 — Jun 2025
Institute of Computing Technology, Chinese Academy of Sciences
Computer Science and Technology
Master of Engineering
Sep 2018 — Jun 2022
Shandong University
Computer Science and Technology
Bachelor of Engineering
Internships
Mar 2025 — Sep 2025
Beijing, China
Honors & awards
- 2025Outstanding Student, State Key Laboratory of Processor Chips, ICT
- 2023, 2024Outstanding Student Scholarship, ICT
- 2021Finalist Award, Mathematical Contest in Modeling (MCM)
- 2021Innovation Scholarship, Shandong University
- 2020Scholarship of Academic Excellence, Shandong University