Effective transfer learning of affordances for household robots

Chang Wang, Koen V. Hindriks, Robert Babuska

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

Abstract

Learning how to use functional objects is essential for robots that are to carry out household tasks. However, learning every object from scratch would be a very naive and time-consuming approach. In this paper, we propose transfer learning of affordances to reduce the number of exploratory actions needed to learn how to use a new object. Through embodied interaction with the object, the robot discovers the object's similarity to previously learned objects by comparing their shape features and spatial relations between object parts. The robot actively selects object parts along with parameterized actions and evaluates the effects on-line. We demonstrate through real-world experiments with the humanoid robot NAO that our method is able to speed up the use of a new type of garbage can by transferring the affordances learned previously for similar garbage cans.

Original languageEnglish
Title of host publicationIEEE ICDL-EPIROB 2014 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages469-475
Number of pages7
ISBN (Electronic)9781479975402
DOIs
Publication statusPublished - 1 Jan 2014
Externally publishedYes
Event4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics, IEEE ICDL-EPIROB 2014 - Genoa, Italy
Duration: 13 Oct 201416 Oct 2014

Conference

Conference4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics, IEEE ICDL-EPIROB 2014
CountryItaly
CityGenoa
Period13/10/1416/10/14

Fingerprint

Robots
Experiments

Cite this

Wang, C., Hindriks, K. V., & Babuska, R. (2014). Effective transfer learning of affordances for household robots. In IEEE ICDL-EPIROB 2014 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics (pp. 469-475). [6983025] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/DEVLRN.2014.6983025
Wang, Chang ; Hindriks, Koen V. ; Babuska, Robert. / Effective transfer learning of affordances for household robots. IEEE ICDL-EPIROB 2014 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 469-475
@inproceedings{b1dfe1177e634f43a796eee000067b62,
title = "Effective transfer learning of affordances for household robots",
abstract = "Learning how to use functional objects is essential for robots that are to carry out household tasks. However, learning every object from scratch would be a very naive and time-consuming approach. In this paper, we propose transfer learning of affordances to reduce the number of exploratory actions needed to learn how to use a new object. Through embodied interaction with the object, the robot discovers the object's similarity to previously learned objects by comparing their shape features and spatial relations between object parts. The robot actively selects object parts along with parameterized actions and evaluates the effects on-line. We demonstrate through real-world experiments with the humanoid robot NAO that our method is able to speed up the use of a new type of garbage can by transferring the affordances learned previously for similar garbage cans.",
author = "Chang Wang and Hindriks, {Koen V.} and Robert Babuska",
year = "2014",
month = "1",
day = "1",
doi = "10.1109/DEVLRN.2014.6983025",
language = "English",
pages = "469--475",
booktitle = "IEEE ICDL-EPIROB 2014 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

Wang, C, Hindriks, KV & Babuska, R 2014, Effective transfer learning of affordances for household robots. in IEEE ICDL-EPIROB 2014 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics., 6983025, Institute of Electrical and Electronics Engineers Inc., pp. 469-475, 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics, IEEE ICDL-EPIROB 2014, Genoa, Italy, 13/10/14. https://doi.org/10.1109/DEVLRN.2014.6983025

Effective transfer learning of affordances for household robots. / Wang, Chang; Hindriks, Koen V.; Babuska, Robert.

IEEE ICDL-EPIROB 2014 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics. Institute of Electrical and Electronics Engineers Inc., 2014. p. 469-475 6983025.

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

TY - GEN

T1 - Effective transfer learning of affordances for household robots

AU - Wang, Chang

AU - Hindriks, Koen V.

AU - Babuska, Robert

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Learning how to use functional objects is essential for robots that are to carry out household tasks. However, learning every object from scratch would be a very naive and time-consuming approach. In this paper, we propose transfer learning of affordances to reduce the number of exploratory actions needed to learn how to use a new object. Through embodied interaction with the object, the robot discovers the object's similarity to previously learned objects by comparing their shape features and spatial relations between object parts. The robot actively selects object parts along with parameterized actions and evaluates the effects on-line. We demonstrate through real-world experiments with the humanoid robot NAO that our method is able to speed up the use of a new type of garbage can by transferring the affordances learned previously for similar garbage cans.

AB - Learning how to use functional objects is essential for robots that are to carry out household tasks. However, learning every object from scratch would be a very naive and time-consuming approach. In this paper, we propose transfer learning of affordances to reduce the number of exploratory actions needed to learn how to use a new object. Through embodied interaction with the object, the robot discovers the object's similarity to previously learned objects by comparing their shape features and spatial relations between object parts. The robot actively selects object parts along with parameterized actions and evaluates the effects on-line. We demonstrate through real-world experiments with the humanoid robot NAO that our method is able to speed up the use of a new type of garbage can by transferring the affordances learned previously for similar garbage cans.

UR - http://www.scopus.com/inward/record.url?scp=84920879886&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84920879886&partnerID=8YFLogxK

U2 - 10.1109/DEVLRN.2014.6983025

DO - 10.1109/DEVLRN.2014.6983025

M3 - Conference contribution

SP - 469

EP - 475

BT - IEEE ICDL-EPIROB 2014 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Wang C, Hindriks KV, Babuska R. Effective transfer learning of affordances for household robots. In IEEE ICDL-EPIROB 2014 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics. Institute of Electrical and Electronics Engineers Inc. 2014. p. 469-475. 6983025 https://doi.org/10.1109/DEVLRN.2014.6983025