CoinDICE: Off-Policy Confidence Interval Estimation via Dual Lens

Channel:
United States Simons Institute for the Theory of Computing
Subscribers:
68,700
Published on ● Video Link: https://www.youtube.com/watch?v=SxfU7hfxwP8


Duration: 25:21
564 views
5

Bo Dai (Google Brain)
https://simons.berkeley.edu/talks/tbd-220
Deep Reinforcement Learning



Other Videos By Simons Institute for the Theory of Computing


2020-10-12Zero-Free Regions for Repulsive Gasses
2020-10-12A Geometric Approach to Conic Stability of Polynomials
2020-10-12Paving Property for Strongly Rayleigh Distributions
2020-10-12Simplicial Generation of Chow Rings of Matroids
2020-10-12Classical Algorithms, Correlation Decay, and Complex Zeros of Quantum Partition Functions
2020-10-12Spectral Sets and Derivatives of The Psd Cone
2020-10-09Robustly Learning Mixtures of (Clusterable) Gaussians via the SoS Proofs to Algorithms Method
2020-10-05Richard M. Karp Distinguished Lecture – Safe Learning in Robotics
2020-10-02Deep Robust Reinforcement Learning and Regularization
2020-10-02Mixed Autonomy Traffic: A Reinforcement Learning Perspective
2020-10-02CoinDICE: Off-Policy Confidence Interval Estimation via Dual Lens
2020-10-02Rigorous Uncertainty Quantification for Off-policy Evaluation in Reinforcement Learning: a Variation
2020-10-02Is Safe Learning the Future of RL?
2020-10-01Policy Gradients Methods, Neural Policy Classes, and Distribution Shift
2020-10-01Fast Reinforcement Learning With Generalized Policy Updates
2020-10-01Exploiting Latent Structure and Bisimulation Metrics for Better Generalization
2020-10-01Invariant Prediction for Generalization in Reinforcement Learning
2020-10-01Language as a Scaffold for Reinforcement Learning
2020-09-30Discussion: Optimization
2020-09-30Generalizing the Projected Bellman Error Objective for Nonlinear Value Estimation
2020-09-30Stabilizing Q-learning with Weighted Bellman Losses


Tags:
Bo Dai
Simons Institute
Theory of Computing
Theory of Computation
Theoretical Computer Science
Computer Science
UC Berkeley
Deep Reinforcement Learning