Computer Science Professor Santosh Vempala was recently named a 2023 Simons Investigator in theoretical computer science by the Simons Foundation.
The Simons Investigator award supports “outstanding theoretical scientists in their most productive years, when they are establishing creative new research directions, providing leadership to the field, and effectively mentoring junior scientists.”
Vempala will receive $150,000 per year for five years from the award, which could be renewed for another five years. Not constrained to any one project, the funding is meant to empower award recipients to push forward on any foundational challenges to computer science (CS) that are related to their interests.
For Vempala, who is the Frederick Storey II Chair of Computing, these challenges have to do with geometry and randomness.
“Some of the most exciting problems in science are those related to efficient algorithms. Computer science really has uncovered fundamentally new questions but also provided novel perspectives on classical problems in mathematics,” Vempala said. “We are yet to understand extremely basic questions such as how best to solve linear systems and linear programs.”
Vempala has been conducting research in theoretical computer science for more than thirty years, and is still finding new interests and perspectives.
“It’s funny, I’m still working on some of the problems I was working on as a Ph.D. student, but certainly some new ones too. For example, I now have an appreciation for non-Euclidean geometry,” Vempala said.
Vempala has recently been researching the brain and building a rigorous theory of how perception and higher-level cognition might emerge from neurons and synapses. He is focused on expanding what this plausible neural model can learn without specialized instructions.
One problem that he has worked on for most of his career is computing the volume of convex bodies. The problem is closely tied to sampling high-dimensional distributions, which has seen much progress over the past decade with radically new techniques introduced by Vempala and others.
“I’d like to understand the limits of this problem, including a lower bound in the general setting, and whether randomness is entirely superfluous in the important case of polytope volume computation,” said Vempala.
Vempala was nominated for the award by Daniel Spielman, a professor at Yale University and a former colleague of Vempala’s.
“Santosh Vempala has done great work his whole career, but his work over the last decade has been particularly amazing,” Spielman said. “He has developed new algorithms for solving important problems and has helped us understand why other problems are so hard to solve. His recent breakthroughs include advances in convex geometry, numerical linear algebra, and computational learning theory.”
School of Computer Science (SCS) Chair Vivek Sarkar said he is excited for Santosh.
"This is a great recognition for Santosh that is not only a testament to his accomplishments but also a positive reflection on Georgia Tech. Simons Investigators include world-class leaders in astrophysics, computer science, mathematics, biology, and physics, placing Santosh in excellent company,” said Sarkar, who is also the Stephen Fleming Chair for Telecommunications in the College of Computing.
Along with his role in SCS, Vempala is also an adjunct professor in the School of Mathematics and the H. Milton Stewart School of Industrial and Systems Engineering. He is also director of Georgia Tech’s Ph.D. program in Algorithms, Combinatorics and Optimization that includes faculty from multiple Georgia Tech schools.
"I am grateful to the Simons Foundation, and to my family, mentors, students, and collaborators,” he said. “I am excited and hopeful about progress on some favorite problems."