The QASE API: A Comprehensive Platform for Games-Based AI Research and Education

Bernard Gorman, Martin Fredriksson, Mark Humphrys

Abstract


Computer games have belatedly come to the fore as a serious platform for AI research. Through our own experiments in the fields of imitation learning and intelligent agents, it became clear that the lack of a unified, powerful yet intuitive API was a serious impediment to the adoption of commercial games in both research and education. Parallel to our own specialised work, we therefore decided to develop a generalpurpose library for the creation of game agents, in the hope that the availability of such software would help stimulate further interest in the field. Though geared towards machinelearning, the API would be flexible enough to facilitate multiple forms of artificial intelligence, making it suitable for application in research and in undergraduate courses centring upon traditional AI and agent-based systems.

In this paper, we present the result of our efforts; the Quake 2 Agent Simulation Environment (QASE) API. We first describe the theme of our work, the reasons for choosing Quake 2 as our testbed, and the necessity for an API of this nature. We then outline its most important features, before presenting an experiment from our own research to demonstrate QASE’s practical capabilities.


Full Text: PDF