Word Based Tic-Tac-Toe Using Reinforcement Learning

This paper introduces a novel approach in playing Tic Tac Toe by replacing traditional Xs and Os with synonyms sourced from a lexical database, WordNet. Leveraging Reinforcement Learning (RL), our agent learns optimal strategies based on the meanings of these words, enhancing gameplay with language understanding. Through the integration of Natural Language ToolKit (NLTK) to download WordNet, and Gensim for importing Word2Vec model, coupled with Q-learning algorithm, our system dynamically adapts its gameplay to the semantic context of the words occupying the game board. This abstracts the game into a domain where linguistic comprehension influences strategic decision-making, presenting an innovative fusion of language and gaming.