SentiMATE: Learning to play Chess through Natural Language Processing
This addresses the problem of sample-efficient chess engine training for AI researchers, though it appears incremental as it builds on existing chess AI methods with a novel NLP integration.
The researchers tackled the problem of training chess engines by developing SentiMATE, which uses natural language processing to learn an evaluation function from chess commentary sentiment, achieving over 90% classification accuracy and beating random agents and a DeepChess implementation at level-one search depth.
We present SentiMATE, a novel end-to-end Deep Learning model for Chess, employing Natural Language Processing that aims to learn an effective evaluation function assessing move quality. This function is pre-trained on the sentiment of commentary associated with the training moves and is used to guide and optimize the agent's game-playing decision making. The contributions of this research are three-fold: we build and put forward both a classifier which extracts commentary describing the quality of Chess moves in vast commentary datasets, and a Sentiment Analysis model trained on Chess commentary to accurately predict the quality of said moves, to then use those predictions to evaluate the optimal next move of a Chess agent. Both classifiers achieve over 90 % classification accuracy. Lastly, we present a Chess engine, SentiMATE, which evaluates Chess moves based on a pre-trained sentiment evaluation function. Our results exhibit strong evidence to support our initial hypothesis - "Can Natural Language Processing be used to train a novel and sample efficient evaluation function in Chess Engines?" - as we integrate our evaluation function into modern Chess engines and play against agents with traditional Chess move evaluation functions, beating both random agents and a DeepChess implementation at a level-one search depth - representing the number of moves a traditional Chess agent (employing the alpha-beta search algorithm) looks ahead in order to evaluate a given chess state.