ViZDoom: A Doom-based AI Research Platform for Visual Reinforcement Learning

TL;DR

ViZDoom is a customizable 3D first-person platform based on Doom, enabling reinforcement learning research with visual inputs, demonstrated by training bots that perform human-like behaviors in complex scenarios.

Contribution

Introduces ViZDoom, a lightweight, customizable 3D platform for visual reinforcement learning from first-person perspectives, bridging the gap between simple 2D games and real-world tasks.

Findings

Bots learned to perform move-and-shoot tasks effectively

Bots demonstrated human-like behaviors in maze navigation

ViZDoom proved useful for visual reinforcement learning research

Abstract

The recent advances in deep neural networks have led to effective vision-based reinforcement learning methods that have been employed to obtain human-level controllers in Atari 2600 games from pixel data. Atari 2600 games, however, do not resemble real-world tasks since they involve non-realistic 2D environments and the third-person perspective. Here, we propose a novel test-bed platform for reinforcement learning research from raw visual information which employs the first-person perspective in a semi-realistic 3D world. The software, called ViZDoom, is based on the classical first-person shooter video game, Doom. It allows developing bots that play the game using the screen buffer. ViZDoom is lightweight, fast, and highly customizable via a convenient mechanism of user scenarios. In the experimental part, we test the environment by trying to learn bots for two scenarios: a basic move-and-shoot task and a more complex maze-navigation problem. Using convolutional deep neural networks with Q-learning and experience replay, for both scenarios, we were able to train competent bots, which exhibit human-like behaviors. The results confirm the utility of ViZDoom as an AI research platform and imply that visual reinforcement learning in 3D realistic first-person perspective environments is feasible.