Vivak Patel
University of Wisconsin-Madison, WI·Mathematics & Statistics
Nearly every large scientific computation, from climate models to engineering simulations, rests on linear algebra: solving systems of equations, factoring matrices, finding dominant patterns in data. At today's data sizes, the classical algorithms become the bottleneck.
- Fields
Mathematics & Statistics
Computer Science
- Topics
randomized numerical linear algebra
matrix sketching
linear algebra algorithms
- Grants
- 1 known grant, through Jun 2027
What they work on
- Map a modular theory of randomized linear algebra algorithms onto composable computational kernels so that new, application-tailored randomized solvers can be assembled rapidly rather than coded from scratch.
- Build a framework for reproducible comparison and pre-deployment benchmarking of randomized linear algebra algorithms, so candidate methods can be evaluated inside real scientific computations before heavy software investment.
- Demonstrate the cyberinfrastructure on large-scale data problems arising in climate applications, using randomized compression of input data to make the computations tractable.
- Develop publicly accessible educational modules, built on the same cyberinfrastructure, covering both the theory and the practical implementation of randomized numerical linear algebra.
More about the research
One promising fix is randomization: instead of touching all the data, an algorithm draws random sketches that compress it while provably preserving what matters. The theory is strong, but the practice lags, largely because there is no shared software infrastructure for building, testing and fairly comparing these randomized methods.
Why it matters
Randomized algorithms could let scientists and engineers tackle data-heavy problems, including climate data analysis, that current linear algebra methods cannot handle at scale. Shared, reproducible software plus benchmarking would let researchers tell which randomized methods genuinely help before investing in development, and the open educational modules make the techniques accessible to newcomers.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Michal Derezinski
University of Michigan, MI·Computer Science·Early-career
Reformulate optimal matrix sketching, the construction of a small randomized approximation of input data, using black-box sampling methods, so that sketch and sub-sample outputs come with reliability guarantees suitable for software.
Atri Rudra
University at Buffalo, NY·Computer Science
Design a family of structured matrices meeting all properties deep learning requires at once, including an efficient projection problem for mapping arbitrary matrices into the family during training.
Fabian Faulstich
Rensselaer Polytechnic Institute, NY·Mathematics & Statistics·Early-career
Recast periodic coupled-cluster theory as a nonlinear Petrov–Galerkin method for a pseudo-differential operator on a high-dimensional torus, in order to derive sharp error estimates for electronic-structure simulations of crystalline materials.
The school
University of Wisconsin-Madison: 91 labs on PI Pages, $1.9B a year on research, 854 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.