Maximum Entropy Dueling Network Architecture in Atari Domain

TL;DR

This paper introduces an enhanced deep reinforcement learning architecture for Atari games that combines Dueling Networks with Maximum Entropy principles, leading to improved policy evaluation.

Contribution

It presents a novel architecture integrating Maximum Entropy with Dueling Networks, enhancing value estimation in Atari domain reinforcement learning.

Findings

Better policy evaluation performance in Atari games

Outperforms original Dueling Networks and other value-based methods

Demonstrates the effectiveness of Maximum Entropy integration

Abstract

In recent years, there have been many deep structures for Reinforcement Learning, mainly for value function estimation and representations. These methods achieved great success in Atari 2600 domain. In this paper, we propose an improved architecture based upon Dueling Networks, in this architecture, there are two separate estimators, one approximate the state value function and the other, state advantage function. This improvement based on Maximum Entropy, shows better policy evaluation compared to the original network and other value-based architectures in Atari domain.