# Can Meta-Interpretive Learning outperform Deep Reinforcement Learning of Evaluable Game strategies?

## TL;DR

This paper compares Meta-Interpretive Learning (MIGO) with standard and deep reinforcement learning on simple games, showing MIGO's superior performance and interpretability in terms of minimax regret and transfer learning.

## Contribution

It introduces MIGO, a new Meta-Interpretive Learning system, and demonstrates its advantages over reinforcement learning in simple game environments with efficient regret evaluation.

## Key findings

- MIGO outperforms reinforcement learning variants in minimax regret.
- MIGO's learned rules are more comprehensible.
- MIGO achieves effective transfer learning between games.

## Abstract

World-class human players have been outperformed in a number of complex two person games (Go, Chess, Checkers) by Deep Reinforcement Learning systems. However, owing to tractability considerations minimax regret of a learning system cannot be evaluated in such games. In this paper we consider simple games (Noughts-and-Crosses and Hexapawn) in which minimax regret can be efficiently evaluated. We use these games to compare Cumulative Minimax Regret for variants of both standard and deep reinforcement learning against two variants of a new Meta-Interpretive Learning system called MIGO. In our experiments all tested variants of both normal and deep reinforcement learning have worse performance (higher cumulative minimax regret) than both variants of MIGO on Noughts-and-Crosses and Hexapawn. Additionally, MIGO's learned rules are relatively easy to comprehend, and are demonstrated to achieve significant transfer learning in both directions between Noughts-and-Crosses and Hexapawn.