Last friday August 28th, 2020 our Head Enrique Zuazua gave the first lecture of the Center for Mathematics and Artificial Intelligence (CMAI) Colloquium this fall about “Turnpike Control and Deep Learning” based on a recent joint work with C. Esteve, B. Geshkovski and D. Pighin.

Introduced by the Director of CMAI Harbir Antil, Enrique Zuazua talked about the turnpike principle asserts that in long time horizons optimal control strategies are nearly of a steady state nature, some recent results on this topic and presented some its consequences on deep supervised learning.

The following video is the complete session recorded: