Towards an Ontology for Describing Emotions 1 st World Summit of the Knowledge Society WSKS08 Juan Miguel López 1, Rosa Gil 1, Roberto García 1, Idoia.

Towards an Ontology for Describing Emotions

1st World Summit of the Knowledge SocietyWSKS’08

Juan Miguel López1, Rosa Gil1, Roberto García1, Idoia Cearreta2, Nestor Garay2

1 Universitat de Lleida, Spain2 University of the Basque Country, Spain

September 25, 2008Athens, Greece

Table of Contents

• Introduction

• Describing Emotion

• Ontologies for Emotion

• Conclusions

• Future Work

• Conceptual Model

• Emotions Ontology

• Use Case

Introduction

• Human beings are eminently emotional• Affective computing: detect and response to

user's emotions• Great variety of theoretical models of

emotions• Emotions are not universal (cultural, language

and individual particularities) Context influence

• Focus (reduce complexity):– Emergent Emotion: states where the person’s whole

system is caught up in the way they react to a particular person or situation

– Just emotion detection and expression systems, not internals of emotion processing in humans

Introduction

• Objectives:– Generic approach to define context-aware

emergent emotions taking different theoretical models into account

– Guide for flexible design of multimodal affective applications with independence of the starting model and the final way of implementation

Describing Emotion

• Most common cognitive models of emotions:– Categorical (Ekman, 1984)– Dimensional (Lang, 1979) – Appraisal (Scherer, 1999)

• Emotion expression systems: – Verbal– Behavioural (e.g. facial or postural)– Psycho-physiological (e.g heart rate)

• Emotional processing levels: – Emotional context

(location, time, activity, devices and person)– Emotion itself– Associated multimodal behaviours

Ontologies for Emotion

• Semantic lexicon in the field of feelings and emotions (Mathieu, 2005)

• Emotional annotation with WordNetAffect (Strapparava and Valittutti, 2004)

• Ontology of affective states for context aware applications (Benta et al., 2007)

• User context model (Cearreta et al., 2007)

Table of Contents

• Introduction

• Describing Emotion

• Ontologies for Emotion

• Conclusions

• Future Work

• Conceptual Model

• Emotions Ontology

• Use Case

Conceptual Model

• Independent from psychological theories– No interpretation of emotions– No emotion triggering mechanism model

• Multimodality:– Incorporates Lang’s three expression systems– Input through senses (humans) and sensors

(computers)

• Model context: individual, social and environmental

• Focus on Emergent Emotion, base of human affectiveness

Conceptual Model

Sensor Sensation

Perception

Description

Stimulus SituationhasPart

triggers

hasInput

hasOutput

describes

Emergent Emotion

Emotion Expression System

hasOutput

hasInput

hasInput

hasOutput

Memorystores

uses

Interface

“physical world” “mental world”

Emotions Ontology

• Formalisation of the conceptual model• Flexible and extensible

(accommodate different theories)• Web-wide sharable:

Web Ontology Language (OWL)• Enrich by reusing upper ontologies• DOLCE, Descriptive Ontology for

Linguistic and Cognitive Engineering (Gangemi et al., 2002)– Context representation: Description & Situation– Other generic concepts

Emotions Ontology

dul:Action

Behavioural

Psychophysiological

Hearing

dul:Event

EmergentEmotion

Stimulus

dul:Process

dul:PhysicalObject

dul:PhysicalAgent

Sensor

dul:PhysicalBody

dul:PhysicalArtifact

Touch

SocialContext dul:Situation

BiologicalSensor Sense

Verbal

DiscreteStimulus

dul:Description ofntb:Frame

Taste

Gestural

Sight

Smell

Sensation

ArtificialSensor

EmotionExpressionSystem

Perception

PersonalContextInterface

Speech

Facial

EnvironmentalContext

ContinuousStimulus

Memory

dul:BiologicalObject

dul:SocialObject

Emotions Ontology

• DOLCE provides generic terms for modelling context

• Enormous range of situations that might be associated with emotions

• FrameNet: formalisation of a enormous linguistic base, based on Frames:

Lexical Unit Frame LU StatusLexical EntryReport

AnnotationReport

score.n Cardinal_numbers Created LE  

score.nBehind_the_scenes

Finished_Initial LE Anno

score.v Getting Created    

score.v Damaging Created LE  

scores.n Quantity Finished_Initial LE Anno

Emotions OntologyScenario

"Torres scored a winning goal in the last minute"

Description

score - Recipient "Torres"- Result "winning"- Theme "goal"- Time "in the last minute"

describes

triggers

Emergent Emotion

Use Case

• Emotion-aware Tangible User Interface• Interface:

– Sensors: microphone, camera and buttons– Expression: display and speaker

• Situations Descriptions: – “playing a song”– “displaying a picture”

Use Case

• Emergent Emotion: sadness, happiness, anger, calm, worry, relaxed, boredom and surprise

• Training: recognize user emotional response to some situations

• Then, make user experience more pleasant– If detected sadness

play songs and/or display images associated to a happy user response

Table of Contents

• Introduction

• Describing Emotion

• Ontologies for Emotion

• Conclusions

• Future Work

• Conceptual Model

• Emotions Ontology

• Use Case

Conclusions

• Generic model for describing emotions and their detection and expression systems taking contextual and multimodal elements into account– Cognitive interpretation of emotions– Independence from emotion theories

• Formalised as a Web Ontology• Reuse DOLCE and FrameNet

Future Work

• Extending the ontology beyond emergent emotion– Affective states and emotions in social

networks

• Extend emotion-aware application based on Tangible User Interfaces

• Make computers more accessible, personalised and adapted to user needs

Thank you for your attention

Roberto Garcí[email protected]