A continuous information gain measure to find the most discriminatory problems for AI benchmarking

Tools

Stephenson, Matthew and Anderson, Damien and Khalifa, Ahmed and Levine, John and Renz, Jochen and Togelius, Julian and Salge, Christoph; (2020) A continuous information gain measure to find the most discriminatory problems for AI benchmarking. In: 2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings. IEEE, GBR. ISBN 9781728169293 (https://doi.org/10.1109/CEC48606.2020.9185834)

[thumbnail of Stephenson-etal-IEEE-CEC-2020-A-continuous-information-gain-measure-to-find]
Preview
 Text. Filename: Stephenson_etal_IEEE_CEC_2020_A_continuous_information_gain_measure_to_find.pdf
Accepted Author Manuscript
License: Strathprints license 1.0
Download (773kB)| Preview

Abstract

This paper introduces an information-theoretic method for selecting a subset of problems which gives the most information about a group of problem-solving algorithms. This method was tested on the games in the General Video Game AI (GVGAI) framework, allowing us to identify a smaller set of games that still gives a large amount of information about the abilities of different game-playing agents. This approach can be used to make agent testing more efficient. We can achieve almost as good discriminatory accuracy when testing on only a handful of games as when testing on more than a hundred games, something which is often computationally infeasible. Furthermore, this method can be extended to study the dimensions of the effective variance in game design between these games, allowing us to identify which games differentiate between agents in the most complementary ways.

ORCID iDs

Stephenson, Matthew, Anderson, Damien ORCID logoORCID: https://orcid.org/0000-0002-8554-3068, Khalifa, Ahmed, Levine, John ORCID logoORCID: https://orcid.org/0000-0001-7016-2978, Renz, Jochen, Togelius, Julian and Salge, Christoph;
CORE (COnnecting REpositories)