Work

Self-organization of action hierarchy and compositionality by reinforcement learning with recurrent neural networks

Year: 2020

Type: article

Abstract: Recurrent neural networks (RNNs) for reinforcement learning (RL) have shown distinct advantages, e.g., solving memory-dependent tasks and meta-learning. However, little effort has been spent on improving RNN architectures and on understanding the underlying neural mechanisms for performance gain. In this paper, we propose a novel, multiple-timescale, stochastic RNN for RL. Empirical results show that the network can autonomously learn to abstract sub-goals and can self-develop an action hierarchy using internal dynamics in a challenging continuous control task. Furthermore, we show that the self-developed compositionality of the network enhances faster re-learning when adapting to a new task that is a re-composition of previously learned sub-goals, than when starting from scratch. We also found that improved performance can be achieved when neural activities are subject to stochastic rather than deterministic dynamics.

Source: Neural Networks

DOI: 10.1016/j.neunet.2020.06.002

Authors: Dongqi Han, Kenji Doya, Jun Tani

Institution: Okinawa Institute of Science and Technology Graduate University

Language: English

Topic: Reinforcement Learning in Robotics

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Estimated APC paid: $3,350

Funders: Okinawa Institute of Science and Technology Graduate University, Japan Society for the Promotion of Science