BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Provably-Correct Neurosymbolic Controllers for Autonomous Cyber-Ph
 ysical Systems - Yasser Soukry
DTSTART:20220727T133000Z
DTEND:20220727T140000Z
UID:TALK176999@talks.cam.ac.uk
DESCRIPTION:While conventional reinforcement learning focuses on designing
  agents that can perform one task\, meta-learning aims\, instead\, to solv
 e the problem of designing agents that can generalize to different tasks (
 e.g.\, environments\, obstacles\, and goals) that were not considered duri
 ng the design or the training of these agents. In this spirit\, we conside
 r the problem of training a provably safe Neural Network (NN) controller f
 or uncertain nonlinear dynamical systems that can generalize to new tasks 
 that were not present in the training data while preserving strong safety 
 guarantees. I will present two complementary neurosymbolic approaches. In 
 the first approach\, I will show how to use ideas from symbolic control to
  provide guarantees on the training of NN controllers. In the second app
 roach\, I will show how to use NN to guide the design of symbolic controll
 ers. I will discuss the theoretical guarantees governing the correctness
  and optimality of these neurosymbolic controllers and show experimental v
 alidation of our approach.
LOCATION:Discussion Room\, Newton Institute
END:VEVENT
END:VCALENDAR