Karthik Elamvazhuthi, Ph.D.
Control theory and Measure Transport Researcher*
New: Notes on Generative Modeling for Control
Research Interests
Multi-robot Control, Optimal Transport, Machine Learning, Generative modeling, Mean Field Control, Geometric Deep Learning
About me
I am fascinated by questions such as how do bees use local feedback to achieve global order and stability without getting at each other’s throats or high stakes squabbling? Can they ever effectively collaborate across scale to provably converge to a peppy title for an academic paper? Would they ever call their highly sophisticated algorithm with sharp convergence guarantees WASP? Are all my worldly non-convex maladies secretly an infinite dimensional linear programming problem in a lifted space of probability measures?
I muse about questions like these to address theoretical and computational aspects of collective behavior in multi-robot systems and machine learning, blending tools from geometric control theory, partial differential equations and optimal transport theory. I do agree that there is no substitute for good theory, except the mountain air and healthy breaks between talks and meetings.
*Funded by coffee, curiosity, and quiet weekend hikes.