Provably adaptive reinforcement learning in metric spaces

06/18/2020
by   Tongyi Cao, et al.
0

We study reinforcement learning in continuous state and action spaces endowed with a metric. We provide a refined analysis of the algorithm of Sinclair, Banerjee, and Yu (2019) and show that its regret scales with the zooming dimension of the instance. This parameter, which originates in the bandit literature, captures the size of the subsets of near optimal actions and is always smaller than the covering dimension used in previous analyses. As such, our results are the first provably adaptive guarantees for reinforcement learning in metric spaces.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
03/09/2020

Zooming for Efficient Model-Free Reinforcement Learning in Metric Spaces

Despite the wealth of research into provably efficient reinforcement lea...
research
05/05/2019

Learning to Control in Metric Space with Optimal Regret

We study online reinforcement learning for finite-horizon deterministic ...
research
07/06/2021

A Short Note on the Relationship of Information Gain and Eluder Dimension

Eluder dimension and information gain are two widely used methods of com...
research
07/07/2020

Sharp Analysis of Smoothed Bellman Error Embedding

The Smoothed Bellman Error Embedding algorithm <cit.>, known as SBEED, w...
research
05/24/2023

Replicable Reinforcement Learning

The replicability crisis in the social, behavioral, and data sciences ha...
research
02/08/2020

Multi-task Reinforcement Learning with a Planning Quasi-Metric

We introduce a new reinforcement learning approach combining a planning ...
research
03/15/2023

Smoothed Q-learning

In Reinforcement Learning the Q-learning algorithm provably converges to...

Please sign up or login with your details

Forgot password? Click here to reset