Our paper "Learning Inverse Statics Models Efficiently with Symmetry-Based Exploration" by R. Rayyes, D. Kubus and J. Steil has been accepted for Frontiers in Neurorobotics journal, research topic: Intrinsically Motivated Open-Ended Learning in Autonomous Robots.
This paper is a breakthrough in the efficiency of inverse models learning. We propose "Symmetry-based exploration" strategy to learn inverse statics models which is valid for the entire configuration space by exploring only a small part of the space and exploiting the symmetry properties of the mapping. Moreover, Goal Babbling and Direction Sampling are modified to learn the inverse statics online, from scratch and in a plain exploration fashion while learning inverse statics models has been done so far only offline.