Exploring Embodiment and Dueling Bandits for Preference Adaptation in Human-Robot Interaction

Sebastian Schneider; Franz Kummert

doi:10.1109/ROMAN.2017.8172476

Exploring Embodiment and Dueling Bandits for Preference Adaptation in Human-Robot Interaction

Sebastian Schneider
, Franz Kummert

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

5 Citations (Scopus)

Abstract

Adaptation for social companions is a crucial requirement for future applications. Personalized interaction seems to be an important factor for long-term commitment to interact with a social robot. We present a study evaluating the feasibility of a dueling bandit learning approach for preference learning (PL) in Human-Robot Interaction (HRI). Furthermore, we explore whether the embodiment of the PL agent has an influence on the user's evaluation of the learner. We conducted a study (n=53) comparing a graphical user interface (GUI), a virtual robot and a real robot. We found no difference regarding the preference for the virtual or real robot. We used the obtained study data to compare the PL approach against a strategy that randomly selects preference rankings. The results show that that the dueling bandit PL approach can be used to learn a user's preference in HRI.

Original language	English
Title of host publication	Human-robot collaboration and human assistance for an improved quality of life
Publisher	IEEE
Pages	1325-1331
Number of pages	7
ISBN (Electronic)	978-1-5386-3518-6, 978-1-5386-3517-9
ISBN (Print)	978-1-5386-3519-3
DOIs	https://doi.org/10.1109/ROMAN.2017.8172476
Publication status	Published - 14 Dec 2017
Externally published	Yes
Event	2017 26th IEEE International Symposium on Robot and Human Interactive Communication - Pestana Palace Hotel, Lisbon, Portugal Duration: 28 Aug 2017 → 1 Sept 2017 Conference number: 26 http://www.ro-man2017.org/site/

Publication series

Name	International Symposium on Robot and Human Interactive Communication
Publisher	IEEE
ISSN (Print)	1944-9437

Conference

Conference	2017 26th IEEE International Symposium on Robot and Human Interactive Communication
Abbreviated title	RO-MAN 2017
Country/Territory	Portugal
City	Lisbon
Period	28/08/17 → 1/09/17
Internet address	http://www.ro-man2017.org/site/

Keywords

Learning (artificial intelligence)
Robot sensing systems
Algorithm design and analysis
Human-robot interaction
Graphical user interfaces
n/a OA procedure

Access to Document

10.1109/ROMAN.2017.8172476

Cite this

@inproceedings{edd86f95d8e242eba6c14f8dd8f2057e,

title = "Exploring Embodiment and Dueling Bandits for Preference Adaptation in Human-Robot Interaction",

abstract = "Adaptation for social companions is a crucial requirement for future applications. Personalized interaction seems to be an important factor for long-term commitment to interact with a social robot. We present a study evaluating the feasibility of a dueling bandit learning approach for preference learning (PL) in Human-Robot Interaction (HRI). Furthermore, we explore whether the embodiment of the PL agent has an influence on the user's evaluation of the learner. We conducted a study (n=53) comparing a graphical user interface (GUI), a virtual robot and a real robot. We found no difference regarding the preference for the virtual or real robot. We used the obtained study data to compare the PL approach against a strategy that randomly selects preference rankings. The results show that that the dueling bandit PL approach can be used to learn a user's preference in HRI.",

keywords = "Learning (artificial intelligence), Robot sensing systems, Algorithm design and analysis, Human-robot interaction, Graphical user interfaces, n/a OA procedure",

author = "Sebastian Schneider and Franz Kummert",

year = "2017",

month = dec,

day = "14",

doi = "10.1109/ROMAN.2017.8172476",

language = "English",

isbn = "978-1-5386-3519-3",

series = "International Symposium on Robot and Human Interactive Communication",

publisher = "IEEE",

pages = "1325--1331",

booktitle = "Human-robot collaboration and human assistance for an improved quality of life",

address = "United States",

note = "2017 26th IEEE International Symposium on Robot and Human Interactive Communication, RO-MAN 2017 ; Conference date: 28-08-2017 Through 01-09-2017",

url = "http://www.ro-man2017.org/site/",

}

Schneider, S & Kummert, F 2017, Exploring Embodiment and Dueling Bandits for Preference Adaptation in Human-Robot Interaction. in Human-robot collaboration and human assistance for an improved quality of life. International Symposium on Robot and Human Interactive Communication, IEEE, pp. 1325-1331, 2017 26th IEEE International Symposium on Robot and Human Interactive Communication, Lisbon, Portugal, 28/08/17. https://doi.org/10.1109/ROMAN.2017.8172476

Exploring Embodiment and Dueling Bandits for Preference Adaptation in Human-Robot Interaction. / Schneider, Sebastian; Kummert, Franz.
Human-robot collaboration and human assistance for an improved quality of life. IEEE, 2017. p. 1325-1331 (International Symposium on Robot and Human Interactive Communication).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Exploring Embodiment and Dueling Bandits for Preference Adaptation in Human-Robot Interaction

AU - Schneider, Sebastian

AU - Kummert, Franz

N1 - Conference code: 26

PY - 2017/12/14

Y1 - 2017/12/14

N2 - Adaptation for social companions is a crucial requirement for future applications. Personalized interaction seems to be an important factor for long-term commitment to interact with a social robot. We present a study evaluating the feasibility of a dueling bandit learning approach for preference learning (PL) in Human-Robot Interaction (HRI). Furthermore, we explore whether the embodiment of the PL agent has an influence on the user's evaluation of the learner. We conducted a study (n=53) comparing a graphical user interface (GUI), a virtual robot and a real robot. We found no difference regarding the preference for the virtual or real robot. We used the obtained study data to compare the PL approach against a strategy that randomly selects preference rankings. The results show that that the dueling bandit PL approach can be used to learn a user's preference in HRI.

AB - Adaptation for social companions is a crucial requirement for future applications. Personalized interaction seems to be an important factor for long-term commitment to interact with a social robot. We present a study evaluating the feasibility of a dueling bandit learning approach for preference learning (PL) in Human-Robot Interaction (HRI). Furthermore, we explore whether the embodiment of the PL agent has an influence on the user's evaluation of the learner. We conducted a study (n=53) comparing a graphical user interface (GUI), a virtual robot and a real robot. We found no difference regarding the preference for the virtual or real robot. We used the obtained study data to compare the PL approach against a strategy that randomly selects preference rankings. The results show that that the dueling bandit PL approach can be used to learn a user's preference in HRI.

KW - Learning (artificial intelligence)

KW - Robot sensing systems

KW - Algorithm design and analysis

KW - Human-robot interaction

KW - Graphical user interfaces

KW - n/a OA procedure

UR - https://pub.uni-bielefeld.de/publication/2909463

U2 - 10.1109/ROMAN.2017.8172476

DO - 10.1109/ROMAN.2017.8172476

M3 - Conference contribution

SN - 978-1-5386-3519-3

T3 - International Symposium on Robot and Human Interactive Communication

SP - 1325

EP - 1331

BT - Human-robot collaboration and human assistance for an improved quality of life

PB - IEEE

T2 - 2017 26th IEEE International Symposium on Robot and Human Interactive Communication

Y2 - 28 August 2017 through 1 September 2017

ER -

Exploring Embodiment and Dueling Bandits for Preference Adaptation in Human-Robot Interaction

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this