Automatic Strategy Inference for Games

We develop algorithms that automatically learn how to play board games. We develop machine learning methods that observe gameplay, and automatically learn strategies from gameplay, pretty much how humans would.

Screenshot of visual algorithm for gameplay

 

Goals:
Automatic machine gameplay for strategy games

Key elements:
Machine learning, Neural networks, LLMs, Large datasets, Python

Research issues:
In this project, we will develop algorithms that automatically learn how to play board games. Here we will develop machine learning methods that observe gameplay, and automatically learn strategies from gameplay, pretty much how humans would. To this end, we will learn about machine learning, neural networks and Large Language Models (LLMs). We will be programming in python and you will also learn about toolsets developed for machine learning applications.

Meeting time:
TBD

Advisors:
Narayana Santhanam, Igor Molybog, and Liuwan Zhu

Partners and Sponsors:
National Science Foundation

Majors, preparation, interests:
geared towards SDS and Comp Eng students, but any Engineering/Math students with an interest in machine learning, neural networks, and artificial intelligence are welcome

Contact information:
Narayana Santhanam, nsanthan@hawaii.edu; Igor Molybog, molybog@hawaii.edu; Liuwan Zhu, liuwan@hawaii.edu