What if we Reduce the Memory of an Artificial Doom Player?

Théo Jaunet, Romain Vuillemot, Christian Wolf    


We built a Doom player AI using Deep Reinforcement learning. While playing, it builds and updates an inner representation (memory) of what it sees from the game. This memory represents what the AI knows about the game, and is the root of each decision. Reducing the size of the memory , could help the player learning to complete its task and thus lower its training time and energy consumption footprint. In this scenario, the player has to gather items in a specific order: Green Armor Red Armor Health Pack Soul-sphere , with the shortest path possible.

@inproceedings{Jaunet:2019,author = {Jaunet, Theo, and  Vuillemot, Romain and  Wolf, Christian},
title = {What if we Reduce the Memory of an Artificial Doom Player?},
journal = {Proceedings of the Workshop on Visualization for AI explainability (VISxAI)},
year = {2019},
editors = {Mennatallah El-Assady, Duen Horng (Polo) Chau, Fred Hohman, Adam Perer, Hendrik Strobelt, Fernanda Viégas},
url = {https://theo-jaunet.github.io/MemoryReduction/}