An iterative algorithm for forward-parameterized skill discovery

Adrien Matricon, David Filliat, Pierre Yves Oudeyer

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

We introduce COCOTTE (COnstrained Complexity Optimization Through iTerative merging of Experts), an iterative algorithm for discovering discrete, meaningful parameterized skills and learning explicit models of them from a set of behaviour examples. We show that forward-parameterized skills can be seen as smooth components of a locally smooth function and, framing the problem as the constrained minimization of a complexity measure, we propose an iterative algorithm to discover them. This algorithm fits well in the developmental robotics framework, as it does not require any external definition of a parameterized task, but discovers skills parameterized by the action from data. An application of our method to a simulated setup featuring a robotic arm interacting with an object is shown.

Original languageEnglish
Title of host publication2016 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages186-192
Number of pages7
ISBN (Electronic)9781509050697
DOIs
Publication statusPublished - 7 Feb 2017
Event2016 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2016 - Cergy-Pontoise, France
Duration: 19 Sept 201622 Sept 2016

Publication series

Name2016 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2016

Conference

Conference2016 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2016
Country/TerritoryFrance
CityCergy-Pontoise
Period19/09/1622/09/16

Fingerprint

Dive into the research topics of 'An iterative algorithm for forward-parameterized skill discovery'. Together they form a unique fingerprint.

Cite this