Approximating two value functions instead of one: towards characterizing a new family of Deep Reinforcement Learning algorithms
Abstract
This paper makes one step forward towards characterizing a new family of \textit{model-free} Deep Reinforcement Learning (DRL) algorithms. The aim of these algorithms is to jointly learn an approximation of the state-value function (), alongside an approximation of the state-action value function (). Our analysis starts with a thorough study of the Deep Quality-Value Learning (DQV) algorithm, a DRL algorithm which has been shown to outperform popular techniques such as Deep-Q-Learning (DQN) and Double-Deep-Q-Learning (DDQN) \cite{sabatelli2018deep}. Intending to investigate why DQV's learning dynamics allow this algorithm to perform so well, we formulate a set of research questions which help us characterize a new family of DRL algorithms. Among our results, we present some specific cases in which DQV's performance can get harmed and introduce a novel \textit{off-policy} DRL algorithm, called DQV-Max, which can outperform DQV. We then study the behavior of the and functions that are learned by DQV and DQV-Max and show that both algorithms might perform so well on several DRL test-beds because they are less prone to suffer from the overestimation bias of the function.
Keywords
Cite
@article{arxiv.1909.01779,
title = {Approximating two value functions instead of one: towards characterizing a new family of Deep Reinforcement Learning algorithms},
author = {Matthia Sabatelli and Gilles Louppe and Pierre Geurts and Marco A. Wiering},
journal= {arXiv preprint arXiv:1909.01779},
year = {2019}
}