Augmenting Automated Game Testing with Deep Reinforcement Learning
Machine Learning
General game testing relies on the use of human playtesters, playtest scripting, and prior knowledge of areas of interest to produce relevant test data. However, by using deep reinforcement learning (DRL), the authors of this paper introduce a self-learning mechanism to the game testing framework.
With DRL, the framework is capable of exploring and exploiting the game mechanics based on a user-defined, reinforcing reward signal. As a result, test coverage is increased and unintended gameplay mechanics, exploits, and bugs can be discovered in a multitude of game types.
In this paper, the authors show that DRL can be used to increase test coverage, find exploits, test map difficulty, and to detect common problems that arise in the testing of first-person shooter (FPS) games.
Presentation at CoG 2020
This paper was presented at the 2020 IEEE Conference on Games (CoG) on August 24–27, 2020. For more information about CoG 2020, visit https://ieee-cog.org/2020/.
Authors: Joakim Bergdahl, Camilo Gordillo, Konrad Tollmar, Linus Gisslén
Watch the presentation below.
The presentation slides are available to download (PDF 26.2 MB).
Download the Paper "Augmenting Automated Game Testing with Deep Reinforcement Learning"
Download the paper as PDF (9.4 MB).