A Continuous Information Gain Measure to Find the Most Discriminatory Problems for AI Benchmarking
Preprint
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.Keywords
All Related Versions
This publication has 6 references indexed in Scilit:
- General Video Game AI: A Multitrack Framework for Evaluating Agents, Games, and Content Generation AlgorithmsIEEE Transactions on Games, 2019
- Dual Indicators to Analyze AI Benchmarks: Difficulty, Discrimination, Ability, and GeneralityIEEE Transactions on Games, 2018
- The 2014 General Video Game Playing CompetitionIEEE Transactions on Computational Intelligence and AI in Games, 2015
- The Arcade Learning Environment: An Evaluation Platform for General AgentsJournal of Artificial Intelligence Research, 2013
- May the Best Analyst Win2011
- General Video Game Playing