On Improving Deep Reinforcement Learning for POMDPs
Abstract
Deep Reinforcement Learning (RL) recently emerged as one of the most competitive approaches for learning in sequential decision making problems with fully observable environments, e.g., computer Go. However, very little work has been done in deep RL to handle partially observable environments. We propose a new architecture called Action-specific Deep Recurrent Q-Network (ADRQN) to enhance learning performance in partially observable domains. Actions are encoded by a fully connected layer and coupled with a convolutional observation to form an action-observation pair. The time series of action-observation pairs are then integrated by an LSTM layer that learns latent states based on which a fully connected layer computes Q-values as in conventional Deep Q-Networks (DQNs). We demonstrate the effectiveness of our new architecture in several partially observable domains, including flickering Atari games.
Cite
@article{arxiv.1804.06309,
title = {On Improving Deep Reinforcement Learning for POMDPs},
author = {Pengfei Zhu and Xin Li and Pascal Poupart and Guanghui Miao},
journal= {arXiv preprint arXiv:1804.06309},
year = {2018}
}
Comments
We are the authors of "On Improving Deep Reinforcement Learning for POMDPs", identifier of which is arXiv:1704.07978. Last week, I wanted to update the article with new version but created a new submission which identifier is 1804.06309 by mistake