Explore the challenges and potential solutions in merging machine learning with control systems for high-dimensional sensor feedback in this 55-minute conference talk. Delve into the pressing research issues surrounding the integration of rich perceptual data, such as camera and microphone inputs, into decision-making systems. Focus on autonomous vehicle control using vision alone as a case study to understand the complexities involved. Learn about innovative approaches to quantifying uncertainty in perception systems, designing robust controllers that account for this uncertainty, and ensuring performance guarantees. Gain insights into the future of control systems that can effectively process and respond to complex sensory inputs, as presented by Ben Recht from the University of California, Berkeley at the Intersections between Control, Learning and Optimization 2020 conference hosted by the Institute for Pure & Applied Mathematics at UCLA.
Trying to Make Sense of Control from Pixels
Institute for Pure & Applied Mathematics (IPAM) via YouTube
Overview
Syllabus
Ben Recht: "Trying to Make Sense of Control from Pixels"
Taught by
Institute for Pure & Applied Mathematics (IPAM)