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

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... (read more)

PDF Abstract

Datasets


Introduced in the Paper:

VizDoom
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Game of Doom ViZDoom Basic Scenario DQN Average Score 82.2 # 1

Methods used in the Paper


METHOD TYPE
Q-Learning
Off-Policy TD Control