Skip to Main content Skip to Navigation
Reports

Dynamic Programming in Distributional Reinforcement Learning

Abstract : The classic approach to reinforcement learning is limited in that it only predicts the expected return. The specialized literature has long tried to remedy this problem by studying risk-sensitive models, but the distributional approach will not emerge until 2017. Since the seminal article M. G. Bellemare, Dabney, and Munos 2017 and the state-of-the-art performance of the C51 algorithm in the ATARI 2600 suite of benchmark tasks (M. G. Bellemare, Naddaf, et al. 2013), research has focused on understanding the behaviour of distributional algorithms. In this paper we place Bellemare's original results in distributional dynamic programming in parallel with the classic results.
Document type :
Reports
Complete list of metadata

https://hal.archives-ouvertes.fr/hal-03168889
Contributor : Elie Odin <>
Submitted on : Sunday, March 14, 2021 - 8:56:25 PM
Last modification on : Friday, April 2, 2021 - 3:23:04 AM

File

Dynamic Programming in Distrib...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-03168889, version 1

Collections

Citation

Odin Elie, Charpentier Arthur. Dynamic Programming in Distributional Reinforcement Learning. [Internship report] Université du Québec à Montréal. 2020. ⟨hal-03168889⟩

Share

Metrics

Record views

94

Files downloads

28