Learning Probabilistic Discriminative Models of Grasp Affordances under Limited 
Supervision

Erkan, AN; Kroemer, O; Detry, R; Altun, Y; Piater, J; Peters, J

doi:10.1109/IROS.2010.5650088

Lokale TagsFreigabegeschichteDetailsÜbersicht

Learning Probabilistic Discriminative Models of Grasp Affordances under Limited Supervision

Erkan, A., Kroemer, O., Detry, R., Altun, Y., Piater, J., & Peters, J. (2010). Learning Probabilistic Discriminative Models of Grasp Affordances under Limited Supervision. In 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2010) (pp. 1586-1591). Piscataway, NJ, USA: IEEE.

Item is Freigegeben

einblenden: alle

Basisdaten

ausblenden:

Datensatz-Permalink: https://hdl.handle.net/11858/00-001M-0000-0013-BDEE-6 Versions-Permalink: https://hdl.handle.net/21.11116/0000-0002-7F1B-3

Genre: Konferenzbeitrag

Dateien

einblenden: Dateien

Externe Referenzen

ausblenden:

externe Referenz:
https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5650088 (Verlagsversion) Open Access Status unbekannt

Beschreibung:
-

OA-Status:

Urheber

ausblenden:

Urheber:
Erkan, AN^{1, 2}, Autor
Kroemer, O^{1, 2}, Autor
Detry, R, Autor
Altun, Y^{1, 2}, Autor
Piater, J, Autor
Peters, J^{1, 2}, Autor

Affiliations:
1Department Empirical Inference, Max Planck Institute for Biological Cybernetics, Max Planck Society, ou_1497795
2Max Planck Institute for Biological Cybernetics, Max Planck Society, Spemannstrasse 38, 72076 Tübingen, DE, ou_1497794

Inhalt

ausblenden:

Schlagwörter: -

Zusammenfassung: This paper addresses the problem of learning and efficiently representing discriminative probabilistic models of object-specific grasp affordances particularly when the number of labeled grasps is extremely limited. The proposed method does not require an explicit 3D model but rather learns an implicit manifold on which it defines a probability distribution over grasp affordances. We obtain hypothetical grasp configurations from visual descriptors that are associated with the contours of an object. While these hypothetical configurations are abundant, labeled configurations are very scarce as these are acquired via time-costly experiments carried out by the robot. Kernel logistic regression (KLR) via joint kernel maps is trained to map the hypothesis space of grasps into continuous class-conditional probability values indicating their achievability. We propose a soft-supervised extension of KLR and a framework to combine the merits of semi-supervised and active learning approaches to tackle the scarcity of labeled grasps. Experimental evaluation shows that combining active and semi-supervised learning is favorable in the existence of an oracle. Furthermore, semi-supervised learning outperforms supervised learning, particularly when the labeled data is very limited.

Details

ausblenden:

Sprache(n):

Datum: Erschienen: 2010-10

Publikationsstatus: Erschienen

Seiten: -

Ort, Verlag, Ausgabe: -

Inhaltsverzeichnis: -

Art der Begutachtung: -

Identifikatoren: DOI: 10.1109/IROS.2010.5650088
BibTex Citekey: 6618

Art des Abschluß: -

Veranstaltung

ausblenden:

Titel: 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2010)

Veranstaltungsort: Taipei, Taiwan

Start-/Enddatum: 2010-10-18 - 2010-10-22

Entscheidung

einblenden:

Projektinformation

einblenden:

Quelle 1

ausblenden:

Titel: 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2010)

Genre der Quelle: Konferenzband

Urheber:

Affiliations:

Ort, Verlag, Ausgabe: Piscataway, NJ, USA : IEEE

Seiten: - Band / Heft: - Artikelnummer: - Start- / Endseite: 1586 - 1591 Identifikator: ISBN: 978-1-424-46675-7