Expanding Neuro-Symbolic Artificial Intelligence for Strategic Learning

Dorian Clay, David C. Anastasiu

Abstract: Today, artificial intelligence (AI) solves problems as varied as driving cars to diagnosing diseases. However, it still has significant pitfalls. For example, the most common form of AI today, machine learning, can produce near-certain predictions for some tasks yet still cannot do reasoning for complex problems and are not well explainable. One solution to these problems is combining today’s cutting-edge neural networks with an older idea in AI: symbolic AI. This neuro-symbolic artificial intelligence has already been shown to excel at visual question answering, a task that involves answering a question about something happening in an image. In the future, such algorithms may play a role in making more flexible and intelligent AI, such as for autonomous driving systems. Right now, neuro-symbolic AI has many areas for expansion across a wide range of applications. This paper will seek to apply neuro-symbolic AI to strategic game play where AI agents have the same information as a human player—just an image. This paper develops the Blokboi game as an AI training and testing tool to enable testing abstraction and high-level reasoning. Blokboi pushes an AI agent to learn scene interpretation and strategic planning and provides an environment that tests an AI’s ability to learn compound interpretation and reasoning. This research expands the applications of NSAI, bringing hybrid artificial intelligence increasingly close to real-world scenarios.