Q-Learning in enormous action spaces via amortized approximate maximization
Abstract
Applying Q-learning to high-dimensional or continuous action spaces can be difficult due to the required maximization over the set of possible actions. Motivated by techniques from amortized inference, we replace the expensive maximization over all actions with a maximization over a small subset of possible actions sampled from a learned proposal distribution. The resulting approach, which we dub Amortized Q-learning (AQL), is able to handle discrete, continuous, or hybrid action spaces while maintaining the benefits of Q-learning. Our experiments on continuous control tasks with up to 21 dimensional actions show that AQL outperforms D3PG (Barth-Maron et al, 2018) and QT-Opt (Kalashnikov et al, 2018). Experiments on structured discrete action spaces demonstrate that AQL can efficiently learn good policies in spaces with thousands of discrete actions.
Cite
@article{arxiv.2001.08116,
title = {Q-Learning in enormous action spaces via amortized approximate maximization},
author = {Tom Van de Wiele and David Warde-Farley and Andriy Mnih and Volodymyr Mnih},
journal= {arXiv preprint arXiv:2001.08116},
year = {2020}
}
Comments
A previous version of this work appeared at the Deep Reinforcement Learning Workshop, NeurIPS 2018