research

My research develops the mathematical and statistical foundations of scientific machine learning. Modern machine learning has produced impressive results, yet many methods still lack generalizability, interpretability, and reliability, which are essential for scientific inference and decision making.

Toward scalable and reliable scientific discovery, my work focuses on three core pillars:

1. scalable operator learning,
2. probabilistic and generative modeling, and
3. uncertainty quantification.

In the long term, I aim to build a coherent framework for reliable scientific machine learning by integrating spectral, probabilistic, and information-theoretic principles into algorithms with provable performance guarantees.

For more details, check out the individual page for each pillar.