Document Type

Conference Proceeding

Publication Title

2004 AAAI Fall Symposium on Real-World Reinforcement Learning. Washington, D.C

Version

Author's Final Manuscript

Publication Date

2004

Abstract

This paper describes a method for designing robots to learn self-motivated behaviors rather than externally specified be- haviors. Self-motivation is viewed as an emergent property arising from two competing pressures: the need to accu- rately predict the environment while simultaneously wanting to seek out novelty in the environment. The robot’s inter- nal prediction error is used to generate a reinforcement signal that pushes the robot to focus on areas of high error or nov- elty. A set of experiments are performed on a simulated robot to demonstrate the feasibility of this approach. The simulated robot is based directly on an existing platform and uses pixe- lated blob vision as its primary sensor.

Share

COinS