AI for Science
In this direction, I develop AI methods in collaboration across the natural sciences, aiming at questions that lead to measurement, explanation, or discovery.
The application areas include materials, physics, and biology. Methodologically, the work is intentionally broad: computer vision, reinforcement learning, foundation models, and multimodal learning all appear where they are useful.
- Scientific imaging and vision for materials and physics
- Texture-aware foundation-model adaptation
- Learning-based optimization in simulation-driven science
