YoVDO

Latent State Recovery in Reinforcement Learning - John Langford

Offered By: Institute for Advanced Study via YouTube

Tags

Reinforcement Learning Courses Theoretical Machine Learning Courses

Course Description

Overview

Explore the intricacies of latent state recovery in reinforcement learning through this comprehensive seminar presented by John Langford from Microsoft Research at the Institute for Advanced Study. Delve into core problems and alternative approaches in reinforcement learning, examining an example problem to understand the challenges involved. Learn about the Homer algorithm, state abstraction techniques, and their proofs. Investigate transition matrices, linear dynamics, and combinatorial state spaces. Engage with a Q&A session and analyze counter-examples to deepen your understanding of this complex topic in theoretical machine learning.

Syllabus

Introduction
Reinforcement Learning
Core Problems
Example Problem
Whats Hard
Questions
Alternative approaches
Hard reinforcement problem
Homer algorithm
State abstraction
Proof
QA
Counter Example
Transition Matrix
Linear Dynamics
combinatorial state space


Taught by

Institute for Advanced Study

Related Courses

On the Critic Function of Implicit Generative Models - Arthur Gretton
Institute for Advanced Study via YouTube
Priors for Semantic Variables - Yoshua Bengio
Institute for Advanced Study via YouTube
Instance-Hiding Schemes for Private Distributed Learning
Institute for Advanced Study via YouTube
Learning Probability Distributions - What Can, What Can't Be Done - Shai Ben-David
Institute for Advanced Study via YouTube
Latent Stochastic Differential Equations for Irregularly-Sampled Time Series - David Duvenaud
Institute for Advanced Study via YouTube