Provably adaptive reinforcement learning in metric spaces

NIPS 2020, 2020.

Cited by: 0|Views84
EI
Weibo:
While the Lipschitz contextual bandits setting of Slivkins is a special case of this setup, no existing analysis recovers his adaptive guarantee that scales with the zooming dimension of the problem

Abstract:

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 \emph{zooming dimension} of the instance. This parameter, which originates in the bandit literature, captures the size o...More

Code:

Data:

0
Your rating :
0

 

Tags
Comments