The YZ Lab

Our lab develops novel machine learning methodologies motivated by challenges in biomedical discovery and scientific understanding. We are particularly interested in integrating ideas from geometry, dynamical systems, optimal transport, and representation learning to build mathematically grounded AI systems for complex biological and high-dimensional data.

Our current research focuses include:

• Generative AI: Flow matching, diffusion models, and optimal transport for cellular dynamics and trajectory inference
• Geometric Deep Learning: Diffusion geometry and manifold learning for single-cell and high-dimensional biomedical data
• Signal Processing: Graph signal processing and PDE-inspired methods for molecular and structural biology
• Brain Decoding: Neural representation learning and dynamical analysis of brain activity and neural systems
• Learning Theory: Theoretical foundations of machine learning, including geometry-aware learning and information-theoretic approaches

By bridging modern machine learning with advanced mathematics, our goal is to develop robust, interpretable, and scientifically meaningful AI methods that advance biomedical research and our understanding of complex systems.

We are hiring PhDs, Postdocs, & Visiting Researchers! Contact Us or Join Us if you are interested!