Agent57: Outperforming the Atari Human Benchmark

Code

• github.com/michaelnny/deep_rl_zoo
  pytorch★ 122
• github.com/pocokhc/agent57
  tf★ 45
• github.com/yuta0821/agent57_pytorch
  pytorch★ 39
• github.com/Nkluge-correa/teeny-tiny_castle
  tf★ 25
• github.com/YHL04/agent57
  pytorch★ 8

Abstract

Atari games have been a long-standing benchmark in the reinforcement learning (RL) community for the past decade. This benchmark was proposed to test general competency of RL algorithms. Previous work has achieved good average performance by doing outstandingly well on many games of the set, but very poorly in several of the most challenging games. We propose Agent57, the first deep RL agent that outperforms the standard human benchmark on all 57 Atari games. To achieve this result, we train a neural network which parameterizes a family of policies ranging from very exploratory to purely exploitative. We propose an adaptive mechanism to choose which policy to prioritize throughout the training process. Additionally, we utilize a novel parameterization of the architecture that allows for more consistent and stable learning.

Tasks

Benchmark Results

Reproductions