LGAIMLJun 14, 2018

Implicit Quantile Networks for Distributional Reinforcement Learning

arXiv:1806.06923v1692 citations
Originality Highly original
AI Analysis

This provides a generally applicable and state-of-the-art method for distributional reinforcement learning, enabling risk-sensitive policies in domains like Atari games.

The paper tackled the problem of distributional reinforcement learning by introducing Implicit Quantile Networks, which approximate the full quantile function for state-action return distributions using quantile regression, resulting in improved performance on 57 Atari 2600 games in the ALE.

In this work, we build on recent advances in distributional reinforcement learning to give a generally applicable, flexible, and state-of-the-art distributional variant of DQN. We achieve this by using quantile regression to approximate the full quantile function for the state-action return distribution. By reparameterizing a distribution over the sample space, this yields an implicitly defined return distribution and gives rise to a large class of risk-sensitive policies. We demonstrate improved performance on the 57 Atari 2600 games in the ALE, and use our algorithm's implicitly defined distributions to study the effects of risk-sensitive policies in Atari games.

Code Implementations19 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