LGAIMLOct 5, 2020

Mastering Atari with Discrete World Models

arXiv:2010.02193v41232 citations
Originality Highly original
AI Analysis

This solves the problem of sample-efficient reinforcement learning for complex environments like Atari games and robotics, representing a significant advance rather than an incremental improvement.

The authors tackled the challenge of modeling Atari games accurately enough to derive successful behaviors by introducing DreamerV2, a reinforcement learning agent that learns behaviors purely from predictions in a compact latent space of a world model, achieving human-level performance on 55 Atari tasks and surpassing top single-GPU agents like IQN and Rainbow with the same computational budget.

Intelligent agents need to generalize from past experience to achieve goals in complex environments. World models facilitate such generalization and allow learning behaviors from imagined outcomes to increase sample-efficiency. While learning world models from image inputs has recently become feasible for some tasks, modeling Atari games accurately enough to derive successful behaviors has remained an open challenge for many years. We introduce DreamerV2, a reinforcement learning agent that learns behaviors purely from predictions in the compact latent space of a powerful world model. The world model uses discrete representations and is trained separately from the policy. DreamerV2 constitutes the first agent that achieves human-level performance on the Atari benchmark of 55 tasks by learning behaviors inside a separately trained world model. With the same computational budget and wall-clock time, Dreamer V2 reaches 200M frames and surpasses the final performance of the top single-GPU agents IQN and Rainbow. DreamerV2 is also applicable to tasks with continuous actions, where it learns an accurate world model of a complex humanoid robot and solves stand-up and walking from only pixel inputs.

Code Implementations9 repos

Data from Papers with Code (CC-BY-SA-4.0)

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes