Deep Active Inference for Partially Observable MDPs
Keywords
Loading...
Authors
Issue Date
2020-07-10
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
Deep active inference has been proposed as a scalable approach to per-ception and action that deals with large policy and state spaces. However, current models are limited to fully observable domains. Here, we describe a deep active inference model that can learn successful policies directly from high-dimensional sensory inputs. The deep learning architecture optimizes a variant of the expected free energy and encodes the continuous state representation by means of a varia-tional autoencoder and a state network. We show, in the OpenAI benchmark, that our approach has better performance than deep Q-learning, a state-of-the-art deep reinforcement learning algorithm.
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen