Energy-based models
Learning expressive energy landscapes for structured scientific data and dynamical systems.
Researcher and builder
Learning, dynamics, and scientific impact.
I develop machine learning methods for scientific systems, with an emphasis on learning dynamics, respecting physical structure, and making models useful in the real world.
Current focus
Generative modeling
Building generative models that support discovery, design, and scientific inference.
Physics-informed learning
Combining physical knowledge and data-driven models for robust prediction and control.
Publications
Research across learning and scientific systems.
- Grain-level micromechanical modeling and assessment of fatigue-critical pores using graph neural networksL Loiodice, KS Stopka, Y Sun, G Lin, MD Sangid · 2026
- Innovating Saltwater Intrusion Monitoring: Deep Learning-Driven Sensor Fusion for Soil Salinity and Moisture MeasurementKJI Egbe, GR Venegas, Y Sun, M Ghayoomi, F Han · 2026
- Disentangled Compositional Diffusion for Controllable Scientific Data GenerationN Madhuj, MH Parikh, A Samaddar, Y Sun, S Madireddy, JX Wang · 2026