Gradient Descent-Ascent Provably Converges to Strict Local Minmax Equilibria with a Finite Timescale

Channel:

Subscribers:

68,700

Published on October 26, 2020 7:41:17 PM ● Video Link: https://www.youtube.com/watch?v=od5GPk_ahQ4

Duration: 30:35

687 views

Other Videos By Simons Institute for the Theory of Computing

2020-10-29	A Generalization Bound for Online Variational Inference
2020-10-28	Model-Based Reinforcement Learning with Value-Targeted Regression
2020-10-28	On the Complexity of Learning Good Policies With and Without Rewards
2020-10-28	A Unifying View of Optimism in Episodic Reinforcement Learning
2020-10-27	Competitive Algorithms for Online Control
2020-10-27	The Non-Stochastic Control Framework
2020-10-27	Robust Algorithms for Secretaries and Bandits
2020-10-27	Regret Minimization for Stochastic Shortest Paths
2020-10-27	Pandora's Box with Correlations: Learning and Approximation
2020-10-26	Learning Outcomes in Queueing Systems
2020-10-26	Gradient Descent-Ascent Provably Converges to Strict Local Minmax Equilibria with a Finite Timescale
2020-10-26	Pure Exploration Problems
2020-10-26	Selfish Robustness and Equilibria in Multi-Player Bandits
2020-10-26	Multi-Task Optimal Experiment Design
2020-10-26	Online Multiserver Convex Chasing and Optimization
2020-10-23	Lower Estimates of Marginal Density
2020-10-23	Graph Matching in Correlated Random Graphs via Small Subgraph Statistics
2020-10-23	The (Non-)Concentration of the Chromatic Number
2020-10-23	Log Concavity and Concentration of Measure on the Discrete Hypercube
2020-10-22	Sharp Poincare and log-Sobolev Inequalities for the Switch Chain on Regular Bipartite Graphs
2020-10-22	Rates of Normal Approximation for Typical Weighted Sums

Tags:

Simons Institute

theoretical computer science

UC Berkeley

Computer Science

Theory of Computation

Theory of Computing

Mathematics of Online Decision Making

Lillian Ratliff