Code associated with the publication: What model does MuZero learn?

DOI:10.4121/a88194f4-45a5-41ae-9cda-0116b78473e5.v1
The DOI displayed above is for this specific version of this dataset, which is currently the latest. Newer versions may be published in the future. For a link that will always point to the latest version, please use
DOI: 10.4121/a88194f4-45a5-41ae-9cda-0116b78473e5
Datacite citation style:
He, Jinke; Moerland, Thomas; de Vries, Joery; Oliehoek, Frans (2024): Code associated with the publication: What model does MuZero learn?. Version 1. 4TU.ResearchData. dataset. https://doi.org/10.4121/a88194f4-45a5-41ae-9cda-0116b78473e5.v1
Other citation styles (APA, Harvard, MLA, Vancouver, Chicago, IEEE) available at Datacite

Dataset

This Dataset contains the code associated with the research paper "What model does MuZero learn?" published at ECAI 2024.

Our research aims to study to what extent models learned by MuZero support policy improvement.

The code here contains scripts to evaluate models learned by MuZero.

Two files implement our policy evaluation and improvement experiments with common functionalities implemented in base.py.

The other scripts are for scaling experiments by automatically generating and launching experiment configurations.

To train MuZero agents, we used https://github.com/YeWR/EfficientZero and https://github.com/werner-duvaud/muzero-general.

History

  • 2024-12-11 first online, published, posted

Publisher

4TU.ResearchData

Format

four of the files are .py files. one is .sh file.

Associated peer-reviewed publication

What model does MuZero learn?

Organizations

TU Delft, Faculty of Electrical Engineering, Mathematics and Computer Science, Department of Intelligent Systems

DATA

Files (6)