Selecting appropriate agent responses based on non-content features

Mark ter Maat; Dirk K.J. Heylen

doi:10.1145/1877826.1877836

Selecting appropriate agent responses based on non-content features

Mark ter Maat
, Dirk K.J. Heylen

Research output: Contribution to conference › Paper › peer-review

2 Citations (Scopus)

7 Downloads (Pure)

Abstract

This paper describes work-in-progress on a study to create models of responses of virtual agents that are selected only based on non-content features, such as prosody and facial expressions. From a corpus of human-human interactions, in which one person was playing the part of an agent and the second person a user, we extracted the turns of the user and gave these to annotators. The annotators had to select utterances from a list of phrases in the repertoire of our agent that would be a good response to the user utterance. The corpus is used to train response selection models based on automatically extracted features and on human annotations of the user-turns.

Original language	Undefined
Pages	33-36
Number of pages	4
DOIs	https://doi.org/10.1145/1877826.1877836
Publication status	Published - Oct 2010
Event	3rd International Workshop on Affective Interaction in Natural Environments, AFFINE 2010 - Firenze, Italy Duration: 25 Oct 2010 → 29 Oct 2010

Workshop

Workshop	3rd International Workshop on Affective Interaction in Natural Environments, AFFINE 2010
Period	25/10/10 → 29/10/10
Other	25-29 October 2010

Keywords

HMI-CI: Computational Intelligence
IR-79593
Machine Learning
EC Grant Agreement nr.: FP7/211486
Virtual agents
EWI-21411
HMI-MI: MULTIMODAL INTERACTIONS
Behaviour selection

Access to Document

10.1145/1877826.1877836

http://eprints.eemcs.utwente.nl/secure2/21411/01/Affine.pdf

Cite this

@conference{e947bbef27ed46d586f71fd73f357c7a,

title = "Selecting appropriate agent responses based on non-content features",

abstract = "This paper describes work-in-progress on a study to create models of responses of virtual agents that are selected only based on non-content features, such as prosody and facial expressions. From a corpus of human-human interactions, in which one person was playing the part of an agent and the second person a user, we extracted the turns of the user and gave these to annotators. The annotators had to select utterances from a list of phrases in the repertoire of our agent that would be a good response to the user utterance. The corpus is used to train response selection models based on automatically extracted features and on human annotations of the user-turns.",

keywords = "HMI-CI: Computational Intelligence, IR-79593, Machine Learning, EC Grant Agreement nr.: FP7/211486, Virtual agents, EWI-21411, HMI-MI: MULTIMODAL INTERACTIONS, Behaviour selection",

author = "\{ter Maat\}, Mark and Heylen, \{Dirk K.J.\}",

year = "2010",

month = oct,

doi = "10.1145/1877826.1877836",

language = "Undefined",

pages = "33--36",

note = "3rd International Workshop on Affective Interaction in Natural Environments, AFFINE 2010 ; Conference date: 25-10-2010 Through 29-10-2010",

}

TY - CONF

T1 - Selecting appropriate agent responses based on non-content features

AU - ter Maat, Mark

AU - Heylen, Dirk K.J.

PY - 2010/10

Y1 - 2010/10

N2 - This paper describes work-in-progress on a study to create models of responses of virtual agents that are selected only based on non-content features, such as prosody and facial expressions. From a corpus of human-human interactions, in which one person was playing the part of an agent and the second person a user, we extracted the turns of the user and gave these to annotators. The annotators had to select utterances from a list of phrases in the repertoire of our agent that would be a good response to the user utterance. The corpus is used to train response selection models based on automatically extracted features and on human annotations of the user-turns.

AB - This paper describes work-in-progress on a study to create models of responses of virtual agents that are selected only based on non-content features, such as prosody and facial expressions. From a corpus of human-human interactions, in which one person was playing the part of an agent and the second person a user, we extracted the turns of the user and gave these to annotators. The annotators had to select utterances from a list of phrases in the repertoire of our agent that would be a good response to the user utterance. The corpus is used to train response selection models based on automatically extracted features and on human annotations of the user-turns.

KW - HMI-CI: Computational Intelligence

KW - IR-79593

KW - Machine Learning

KW - EC Grant Agreement nr.: FP7/211486

KW - Virtual agents

KW - EWI-21411

KW - HMI-MI: MULTIMODAL INTERACTIONS

KW - Behaviour selection

U2 - 10.1145/1877826.1877836

DO - 10.1145/1877826.1877836

M3 - Paper

SP - 33

EP - 36

T2 - 3rd International Workshop on Affective Interaction in Natural Environments, AFFINE 2010

Y2 - 25 October 2010 through 29 October 2010

ER -