AMaze: a lightweight benchmark generator for sighted agents

Research output: Web publication or Non-textual formArtefactAcademic

Abstract

AMaze is a library for fast prototyping generic machine learning agents (Reinforcement Learning, Evolutionary Computation, NeuroEvolution, ...). It relies on a lightweight sequential visual task geared towards human interaction with highly customizable and intelligible difficulties. Sources: https://github.com/kgd-al/amaze Documentation: https://amaze.readthedocs.io/
Original languageEnglish
DOIs
Publication statusPublished - 2 Apr 2024

Fingerprint

Dive into the research topics of 'AMaze: a lightweight benchmark generator for sighted agents'. Together they form a unique fingerprint.

Cite this