Suvrit's Research Interests

My research focuses on large-scale data analysis and optimization In particular, I design, analyze, and implement optimization algorithms for large-scale (data intensive) problems in statistics, machine learning, and computational science. My main mathematical tools draw upon: theoretical computer science, statistics, signal processing, matrix analysis, harmonic analysis, convex and nonconvex optimization, stochastic programming, and numerical linear algebra.

More broadly, I am usually interested in all things computational, where one can not only implement algorithms, but also prove some theorems!

Primary Projects and Research Areas

1. Machine learning, data mining, computational statistics
2. Inverse Problems in Signal Processing, Medical Imaging, Astronomy, Computer Vision
3. Inexact Computation: Optimization under error and uncertainty
4. Large-scale (continuous) optimization:
   • Large-scale linear and quadratic programming
   • Distributed and Parallel Nonlinear Optimization
   • Large-scale nonsmooth convex and nonconvex optimization
5. Search and retrieval in databases of structured objects (images, proteins, etc.)
6. Matrix Analysis: matrix means, positive definite matrices, kernel function theory, etc.

Other Projects and Research Areas

1. Computational finance
2. Statistics on Manifolds (hyperspheres, etc.)
3. Numerical analysis with special functions
4. Inverse eigenvalue problems
5. Statistics and optimization problems on large graphs
6. Operator inequalities
7. Noncommutative algebra (nascent interest)