معرفی
Suvrit Sra is the Esther and Harold E. Edgerton Career Development Associate Professor in the Department of Electrical Engineering and Computer Science at the Massachusetts Institute of Technology (MIT). He is a core member of the Institute for Data, Systems, and Society (IDSS) and the Laboratory for Information and Decision Systems (LIDS). Currently, he is on leave at the Technical University of Munich (TU Munich).
His research centers on the mathematics of artificial intelligence, with emphases on optimization for machine learning (especially non-convex, non-Euclidean, and geometric optimization), theory of deep learning, discrete probability, sampling methods, convex geometry, and polynomials. He also explores applications in operations research, supply chains, and large language models.
His recent articles demonstrate a strong focus on Riemannian optimization, convergence analysis of gradient methods, minimax optimization in geometric spaces, and invariant representations in graph learning. Key trends include theoretical advances in non-convex optimization and applications to reinforcement learning and distributional robustness.
Awards and honors include:
- Alexander von Humboldt Professorship for AI (2024)
- Criteo Faculty Research Award (2017)
- SIAM Outstanding Paper Prize (2011)
He advises PhD students (e.g., Zelda Mariet, Chengtao Li, Hongyi Zhang) and has secured grants including an NSF BIGDATA award for interpretable learning via discrete probability. His lab affiliations include the MIT ML Group, Center for Statistics, and TILOS AI Research Institute. He co-founded Pendulum, serving as Chief Scientist.
