CIS Seminar: “Equilibrium Complexity and Deep Learning”
October 18, 2022 at 3:30 PM - 4:30 PM
Share this event
Details
Date:
October 18, 2022
Time:
3:30 PM - 4:30 PM
Organizer
Computer and Information Science
Phone:
215-898-8560
Email:
cherylh@cis.upenn.edu
Website:
View Organizer Website
Venue
Deep Learning has recently made significant progress in learning challenges such as speech and image recognition, automatic translation, and text generation, much of that progress being fueled by the success of gradient descent-based optimization methods in computing local optima of non-convex objectives. From robustifying machine learning models against adversarial attacks to causal inference, training generative models, multi-robot interactions, and learning in strategic environments, many outstanding challenges in Machine Learning lie at its interface with Game Theory. On this front, however, Deep Learning has been less successful. Here, the role of single-objective optimization is played by equilibrium computation, but gradient-descent based methods fail to find equilibria, and even computing local equilibria — the analog of computing local optima in single-agent settings — has remained elusive.
We shed light on these challenges through a combination of learning-theoretic, complexity-theoretic, and game-theoretic techniques, presenting obstacles and opportunities for Machine Learning and Game Theory going forward, including recent progress on multi-agent reinforcement learning.
(I will assume no deep learning, game theory, or complexity theory background for this talk and present results from joint works with Noah Golowich, Stratis Skoulakis, Manolis Zampetakis, and Kaiqing Zhang.)