Emotion Modeling: Towards Pragmatic Computational Models of Affective Processes

171 Pages·2014·10.33 MB·English

by  Tibor Bosse

t Tibor Bosse
r
A
- Joost Broekens
e
h João Dias
t
-
f Janneke van der Zwaan (Eds.)
o
y
-e
e
v
t
ar
u
t
SS
0
5
7
8

AI Emotion Modeling
N
L
Towards Pragmatic Computational Models
of Affective Processes
123
Lecture Notes in Artificial Intelligence 8750
Subseries of Lecture Notes in Computer Science
LNAI Series Editors
Randy Goebel
University of Alberta, Edmonton, Canada
Yuzuru Tanaka
Hokkaido University, Sapporo, Japan
Wolfgang Wahlster
DFKI and Saarland University, Saarbrücken, Germany
LNAI Founding Series Editor
Joerg Siekmann
DFKI and Saarland University, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/1244
Tibor Bosse Joost Broekens
(cid:129)
ã
Jo o Dias Janneke van der Zwaan (Eds.)
(cid:129)
Emotion Modeling
Towards Pragmatic Computational Models
of Affective Processes
123
Editors
TiborBosse João Dias
VUUniversityAmsterdam Universityof Lisbon
Amsterdam PortoSalvo
The Netherlands Portugal
JoostBroekens Jannekevan derZwaan
Delft Universityof Technology Delft Universityof Technology
Delft Delft
The Netherlands The Netherlands
ISSN 0302-9743 ISSN 1611-3349 (electronic)
Lecture Notesin ArtificialIntelligence
ISBN 978-3-319-12972-3 ISBN 978-3-319-12973-0 (eBook)
DOI 10.1007/978-3-319-12973-0
LibraryofCongressControlNumber:2014955538
LNCSSublibrary:SL7–ArtificialIntelligence
SpringerChamHeidelbergNewYorkDordrechtLondon
©SpringerInternationalPublishingSwitzerland2014
Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartofthe
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilarmethodologynow
knownorhereafterdeveloped.
Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication
doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant
protectivelawsandregulationsandthereforefreeforgeneraluse.
Thepublisher,theauthorsandtheeditorsaresafetoassumethattheadviceandinformationinthisbookare
believedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsortheeditors
give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissionsthatmayhavebeenmade.
Printedonacid-freepaper
SpringerInternationalPublishingAGSwitzerlandispartofSpringerScience+Business
Media(www.springer.com)
Preface
As research on emotion modeling is becoming more mature, the amount of compu-
tationalmodelsofaffectiveprocessesisrapidlyincreasing.Nevertheless,itisimportant
to cut down the complexity of these models, for multiple reasons.
First, cutting down complexity facilitates reuse of models. Many computational
models of affective processes are developed primarily for a practical purpose and are
therefore embedded in something else (e.g., an adaptive agent, intelligent virtual
character, or robot). These models are often developed (a) in an ad hoc way, building
on the best practices from earlier research, (b) based on emotional theories that lack
necessary computational details, (c) from scratch, using publications of others instead
of using components from existing models or (d) using assumptions from others, and
finally (e) without a solid validation scheme. This is understandable given the applied
nature of many of these models. However, this approach results in computational
modelsofaffectiveprocessesthatarecomplex,anddifficulttounderstandandvalidate.
A pragmatic approach is easier to understand, build upon, and implement.
Second, limiting complexity facilitates contributions to psychological research.
Computationalmodelsofemotionaretoolsforexperimentalpsychologiststhatcanbe
used to gain insight into emotion. Many of the computational mechanisms needed for
developing these models are not available in emotion psychology. Hence, instead of
askingforthesemechanismsfrompsychologists, themodelsshouldbringinsight,and
generate testable hypotheses. For this to be possible the model must be fully specified
in a form that is understandable for psychologists, and have clear psychological
grounding. Simpler models are easier to understand and ground than complex ones.
The goal of this volume is to gain more insight into the knowledge that is used to
developcomputationalmodelsofaffectiveprocesses.Tothisend,thebookstartswith
an introduction to the field, written by Eva Hudlicka. After that, it contains extended
versionsofeightpapersthatwerepresentedattheworkshoponStandardsforEmotion
Modeling (held in Leiden, the Netherlands, in 2011) and the workshop on Emotional
andEmpathicAgents(heldinValencia,Spain,in2012).Toassureahighquality,each
chapterwasreviewedbyatleastthreeanonymousreferees.Thechaptersaredistributed
over two parts: one part on “Generic Models and Frameworks,” and the other one on
“Evaluations of Specific Models.” Nevertheless, the material of each chapter is self-
contained, so that readers can select any individual chapter based on their research
interests without the need of reading other chapters.
VI Preface
To conclude, we would like to thank all the authors for their interesting contribu-
tions, and in particular Eva Hudlicka for writing an excellent introduction chapter.
Also,weexpressourgratitudetothereviewersfortheirhardworkinassuringthehigh
qualityofthissurvey,andtoAlfredHofmannandAnnaKrameratSpringerVerlagfor
providing us the opportunity to publish this volume.
August 2014 Tibor Bosse
Joost Broekens
João Dias
Janneke van der Zwaan
Contents
Introduction
From Habits to Standards: Towards Systematic Design of Emotion Models
and Affective Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Eva Hudlicka
Generic Models and Frameworks
Creating a World for Socio-Cultural Agents. . . . . . . . . . . . . . . . . . . . . . . . 27
Nick Degens, Gert Jan Hofstede, John Mc Breen, Adrie Beulens,
Samuel Mascarenhas, Nuno Ferreira, Ana Paiva, and Frank Dignum
FAtiMA Modular: Towards an Agent Architecture with a Generic
Appraisal Framework. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
João Dias, Samuel Mascarenhas, and Ana Paiva
Programming Agents with Emotions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Mehdi Dastani, Christiaan Floor, and John-Jules Ch. Meyer
Utilizing Emotions in Autonomous Robots: An Enactive Approach. . . . . . . . 76
Robert Lowe and Kiril Kiryazov
Evaluations of Specific Models
The Effect of Dominance Manipulation on the Perception and Believability
of an Emotional Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Wim F.J. van der Ham, Joost Broekens, and Peter H.M.P. Roelofsma
Modelling Two Emotion Regulation Strategies as Key Features
of Therapeutic Empathy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Juan Martínez-Miranda, Adrián Bresó, and Juan Miguel García-Gómez
Social Support Strategies for Embodied Conversational Agents. . . . . . . . . . . 134
Janneke M. van der Zwaan, Virginia Dignum, and Catholijn M. Jonker
Affective Processes as Network Hubs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
David Gibson
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Introduction
From Habits to Standards: Towards
Systematic Design of Emotion Models
and Affective Architectures
Eva Hudlicka(&)
Psychometrix Associates andUniversity of Massachusetts Amherst,
Amherst, MA, USA
[email protected]
Abstract. Emotionmodelinghasbeenanactiveareaofresearchforalmosttwo
decadesnow.Yetinspiteofthegrowinganddiversebodyofwork,designing
and developing emotion models remains an art, with few standards and sys-
tematic guidelines available to guide the design process, and to validate the
resulting models. In this introduction I first summarize some of the existing
workattemptingtoestablishmoresystematicapproachestoaffectivemodeling,
andhighlightthe specificcontributionstothis effortdiscussed inthe papersin
this volume. I then propose an analytical computational framework that delin-
eates the core affective processes, emotion generation andemotion effects,and
defines the abstract computational tasks necessary to implement these. This
framework provides both a common vocabulary for describing the computa-
tional requirements for affective modeling, and proposes the building blocks
necessary for implementing emotion models. As such, it can serve both as a
foundationfordevelopingmoresystematicguidelinesformodeldesign,andasa
basis for developing modeling tools. I conclude with a summary and a dis-
cussion of some openquestionsandchallenges.
1 Introduction
Emotion modeling has been an active area of research for almost two decades now.
Since 1997, when Picard coined the term affective computing (Picard 1997), emotion
modeling, and affective computing in general, have blossomed. In the broader area of
affective computing there is now a dedicated bi-ennial conference (ACII) and two
journals: the International Journal of Synthetic Emotions and IEEE Transactions on
AffectiveComputing,aswellasnumerousspecialissuesofotherjournalsfocusingon
some aspect of affective computing (e.g., Intl. Jnl. of Human-Computer Studies, User
Modeling and User Adapted Interaction). A number of AAAI symposia have been
dedicated to the exploration of specific aspects of emotion modeling (cognition/emo-
tion interaction, affective architectures, emotions in social agents), and numerous
workshops have been organized in conjunction with larger conferences, focusing on
affectivemodelinginspecificcontexts,includingagents(IVA,AAMAS),user-adapted
interaction (UMUAI), gaming (Foundations of Digital Games), emotion recognition
(Automatic Face and Gesture Recognition/FG), and models of affective disorders
(Cognitive Science).
©SpringerInternationalPublishingSwitzerland2014
T.Bosseetal.(Eds.):EmotionModeling,LNAI8750,pp.3–23,2014.
DOI:10.1007/978-3-319-12973-0_1