Domain-Independent Game Abstraction using Word Embedding Techniques

2026-05-15

AI Summary

Existing game abstraction methods often rely heavily on domain-specific knowledge—such as that from poker—and thus exhibit limited generalizability. This work proposes a domain-agnostic framework for game abstraction by introducing word embedding techniques from natural language processing into the game-theoretic context: game actions are treated as words and sequences of plays as corpora, enabling an embedding model to learn real-valued vector representations of actions. These embeddings are then combined with clustering algorithms to automatically abstract the action space. Experimental results demonstrate that the approach effectively captures structural information inherent in games without requiring any prior domain knowledge. Although its performance currently does not surpass that of specialized algorithms, the method exhibits strong generalization capabilities and broad applicability across diverse game domains.

Abstract

Many games of interest in the real world are often intractably large, thereby necessitating the use of game abstraction to shrink them in size, typically by many magnitudes. Over the last two decades, there have been significant advances in game abstraction; however, the domain-specific nature (usually poker) of much of the prior work prevents those techniques from being easily generalized to other settings without extensively analyzing the game at hand. In this paper, we propose a domain-independent approach to game abstraction, which applies word embedding techniques from the field of natural language processing. Treating each action as a word and gameplay data as a corpus, word vectors can be trained to represent each action as a real-valued vector, which can then be clustered to facilitate game abstraction. We also explore the use of foundational embedding models and show that action embeddings obtained this way can capture a surprising amount of information about the underlying game. Experimental results demonstrate that our proposed game abstraction technique is effective, although it does not outperform specialized algorithms tailored to specific games.

Problem

• game abstraction
• domain-independent
• word embedding
• action representation
• generalization

Innovation

• game abstraction
• word embedding
• domain-independent
• action representation
• clustering