Affective Computing

1
Affective Computing

• Introduction
• Emotions are Physiological and Cognitive Chp. 1
• TR-321

2
Picard - Introduction

• Picards introduction explains why we should
  incorporate emotions into a computer.
• Emotions play a major role in motivation,
  perception, cognition, coping, creativity,
  attention, planning reasoning, learning, memory,
  and decision making.
• a general theory of thinking and problem
  solving must incorporate the influences of
  emotion - Herb Simon

3
Picard Introduction cont

• Affective Computing is computing that relates
  to, arises from, or deliberately influences
  emotions
• It Includes giving a computer the ability to
• Implement Emotions
• Recognize and Express Emotions
• Respond Intelligently to Human Emotions
• Regulate and Utilize its Emotions

4
Picard Introduction cont

• The Limbic System
• The Limbic system is considered the primary seat
  of emotion, attention, and memory.
• It helps determine
• Valence (whether you feel positive
• or negative about something)
• Salience (what gets your attention)

5
Picard Introduction cont

• The Limbic-Cortical Tangle
• incorrect belief the cortex alone handles all
  visual and audio inputs.
• The the limbic and cortical areas of the brain do
  not work in isolation.
• Experiments have shown that the limbic system can
  recognize certain inputs without the help of the
  cortex. (rat experiment)
• Cortex is still required for more complex inputs

6
Picard Introduction cont

• Thinking/Feeling Axis
• The Myers-Briggs personality type indicator
  places thinking and feeling on opposite endpoints
  on an axis.
• Studies show that
• 2/3 men lie closer to thinking
• 2/3 women lie closer to feeling
• Technical graduate students lie closer to thinking

7
Picard Introduction cont

• Turing Test can not distinguish between a human
  or computer with a conversation.
• In a Turing Test, Affective Computing would
  effect the manner in which the computer would
  talk, i.e. angry, happy, etc..
• Affective Computing can also help computers
  become aware of the emotions of others in a
  social environment.

8
Book Review of Picard by Prof. Lisetti

• The pre/postcognitive debates
• Does affect occurs alongside thought or is it the
  result of some mental process.
• The more important issue is how can emotions be
  generated in a computer, recognized by a computer
  and express by a computer.

9
Review of Picard cont

• No single good computational model of affect.
  Instead a model should be geared towards the
  objective at hand.
• For example, emotion recognition and emotion
  expression are very different problems.
• Best model is a hybrid of subsystems

10
Review of Picard cont

• computer should be adapting to people instead of
  vice versa
• We are communicating increasingly with our
  computers in a growing number of contexts.
• Much work to be done in Human-Computer
  Interaction, facial recognition, synthesis of
  affect in speech, and agent animation.

11
Review of Picard cont..

• Emotions
• Adds flexibility of behavior
• Increases ability to efficiently and
  intelligently address multiple concerns
• Helps determine event valence and salience
• Influences low-level perceptual processes
• Coordinates adaptive behavior when challenged
• Effects decision making abilities

12
Review of Picard cont

• Pre 1960s
• James emotions play valuable role
• Cannon emotions are disruptive
• Post 1960s
• Emotion renaissance
• Findings about universality and specificity drove
  this renaissance.
• AI initiation predated(1956) this renaissance

13
Review of Picard cont.

• Overall a positive review of the book
  admirable, ambitious, brilliant, great
• Only really disagreed on one thing
• Picard says subjective emotions are the hardest
  to achieve and Prof. Lisetti said they are
  impossible because of the biological make up of
  human that computers lack.

14
Affective Computing

• Emotions are Physical and Cognitive - Chapter 1

15
How can we study affective computing?(when
emotions are so hard to precisely define)

• McCarthys Mt. Everest

• Might not be defined, but can be
• detected and emulated

• Studying it in computers might
• will help us better understand
• them in ourselves

16
Two Views of Emotion

• Why two views?

• Damasio Descartes Error

• Physical expression of the body

• Cognitive state of mind

17
Physical Emotion

• William James

• Emotions are the physiological expression of an
  exciting fact

• http//www.emory.edu/EDUCATION/mfp/james.html

18
Components of Emotion

• Emotional state internal dynamics

• Emotional experience subjective experience

• Emotional expression external experience of
  emotion by others

• Mood long-term emotional state

19
Sentic Modulations

• Physically experienced emotional expressions.

• Facial change (Ekman, Duchenne)

• Voice inflection

• Body Posture

• Sentic Equivalence (Clynes)

20
Facial Expression

• Duchenne muscles of attention, disdain, lust,
  doubt

• Ekman - Facial Action Coding System

• Social Display Rules

• http//www.paulekman.com/

21
Sentic Equivalence
Dr. Manfred Clynes

• Sentograph two degrees of freedom

• Essentic Form a spatiotemporal form that
  embodies an emotional message

• http//www.superconductor.com/clynes/default.htm

22
Physiological Factors

• Intensity of Emotion

• Type of Emotion

• Cause of Emotion

• Social Display Rules

23
Altering Expression

• Chemical hormones, medication

• Mood states

• Other non-emotional triggers

24
Problems Detecting

• Idiosyncratic Expression

• Ambiguous Sentic Expressions

• Expression Altering Factors

• Wearables Solution (Negroponte)

25
Cognitive Emotions

• Dr. Antonio Damasio

• Primary Emotions - instinctive

• Secondary Emotions - cognitive

• http//www.uiowa.edu/neuro/Faculty/damasioa.htm

26
Secondary Emotions

• Help generalize reactions to primary

• Help with memory (Stanislavsky)

• Dictated by personal experience

27
How Affect Affects You

• Emotion helps the interaction between limbic and
  cortical systems creativity!

• Alcohol and Caffeine

• Whistle while you work!

• Elephants are Happy! -)

• Physiological Feedback

28
Affective Computingby R. W. PicardTechnical
Report No. 321
29
Affective Computing

• Computing that relates to, arises from, or
  deliberately influences emotions.

30
Emotions are not a luxury, they have an
essential role in human cognition

• Rational decision-making
• Perception
• Human Interaction
• Human Intelligence

31
The Limbic-Cortic Tangle

• Limbic system Primary seat of emotion,
  attention, and memory.
• Cortex The majority of perceptual processing has
  been assumed to occur
• Limbic and cortical brain areas do not operate in
  isolation, but are functionally intertwined.

32
Decision making

• Too much emotion impairs decision- making. Too
  little emotion does the same, too.
• Example Damasios patients with frontal-lobe
  disorders

33
Test of thinking

• To pass the Turing test the computer responses
  should be indistinguishable from human responses.
  
• Because humans respond with emotion to highly
  emotional event it is impossible for a computer
  to pass this test if it lacks affect.

34
Physical and Cognitive Aspects

• Being cognitive (mental component)
• - Situations that give rise to emotions.
• Being physical (physical component)
• - Physiological response that co-occurs with
  an emotion or rapidly follows it.

35
Physical Aspects

• Voice inflection
• Facial expression
• Posture
• Heart rate, pulse
• Temperature
• Blood pressure

36
Affective Pattern Recognition

• Computer recognition of affective states
  appears doable via the measurement of sentic
  modulation.
• The problem is to decide what signals to
  measure, how to process these signals and how to
  interpret them.

37
Modeling Affect

• Discrete affective state models
• Continuous affect models
• Cathexis in computing
• Rule-based emotion models

38
A dilemma

• Can we create computers that will recognize
  and express affect, exhibit creativity,
  intelligent problem solving, and empathy, and
  never bring about harm by emotional actions?

39
Categories of Affective Computing

• computers that
• 1. cannot perceive and express affect
• 2. cannot perceive but can express affect
• 3. can perceive but cannot express affect
• 4. can perceive and express affect

40
Applications of Affective Computing

• Entertainment
• Expressive Communication
• Film/Video
• Learning
• Affective Environments
• Aesthetic Pleasure
• Wearable Computers