Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

Man and Machine > Task 7 > Flashcards

Task 7 Flashcards

1
Q

Affect and Mood

A

Affect: underlying experience of feelings and emotions & mood

Mood: more long lasting, less directed towards a certain situation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Affective computing

A

research and development of systems and devices which can identify, interpret, process and respond to humans emotional state

TYPES:

  1. systems that detect emotions of the user
  2. systems that express what humans would perceive as emotion
  3. systems that feel an emotion
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

How does affective computing work

A
  1. Technologies sense emotions of a user by using some combinations of four inputs:
    a) facial recognition
    b) voice recognition
    c) gestures
    d) biometrics
  2. Computer software is then used to analyze the gathered data: estimate users emotional state and provide an appropriate response
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Methods of affective computing

A
  1. knowledge based: informing search processes by databases on emotional words
  2. statistical approach: deep learning
  3. Hybrid approach: uses 1) and 2)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Benefits of affective computing

A

EMOTIONAL INTELLIGENCE: ability of a person to recognize own feelings or emotions and those of others and manage and adjust them to achieve a goal

  • without EI a computing system might become smarter but not more useful
  • AI can detect conversational cues (e.g. rise of pitch in voice)
  • Detect brief facial expressions and subtle emotional cues (help people struggle with hidden issues, such as depression)
  • Analyze facial and audiovisual displays of affective states
  • emotion and pain project: how sensors can read user levels of pain
  • improve empathy: help autistic patients
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Concerns about affective computing

A
  1. Privacy: data collected can convey extremely personal things - guidelines are needed so that the systems can’t be manipulated
  2. Lack of consideration given to culturally based emotions: emotions may not be consistent across cultures, leading to unreliability
  3. costs: equipment needed to achieve good results is costly
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Cyberpsychology

A

emerging discipline focussing on human-machine interaction: how humans are impacted by technology, how they interact online

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Cybercrime

- why do we fall for fraud

A

a) Internet specific crimes: hacking
b) Internet enabled crimes: theft, human trafficking, advance free fraud (minor request – big reward ‘you won 2 mio dollar, only pay 30 cents to call’)

WHY DO WE FALL FOR THESE FRAUDS?
Biases involved:
a) optimism bias: oneself is less likely to experience a negative event
b) instant gratification (involved in piracy): why wait for it on Netflix? I want to watch it now

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Forensic cyber psychology

A

They try to:

a) predict offending
b) enhance understanding of offender
c) determine most appropriate assessment of offender

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Theories of crime

A

Neutralization: develop ways of rationalize behavior to reduce the feelings of guilt (everybody is doing it - Coleman)

Routine Activity Theory: For a crime to happen, there needs to be a likely offender, a suitable target and the absence of guardian

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Study cyber psychology:

A

the greater the perception of threat, the greater the hardline militant attitudes of people
- stress doesn’t play a role in that

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Self and identity in cyberspace

A

SELF: AS essential being
IDENTITY: compared to others with regard to one’s potentials and qualities
VIRTUAL SELF: online without any real life consequences:
- all 3 types of self (ideal, actual and ought) can be expressed and physical gatherings are overcome = identity empowerment

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Anonymity in cyberspace

A

Equalization hypothesis: removal of social cues = less stereotyped = more socialization

Social model of deindividuation:

a) cognitive component: how individuals behave ands see themselves within a group (group completely anonymous = increase in group salience vs. one person anonymous = decrease in group salience)
b) strategic component: to achieve goal directed groups, we favor:
1. complete anonymity 2. complete transparence (each member)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Different behavior in social network sites:

A

a) broadcaster - speaking in front of group, self-focussed, self-promo
b) communicating - group focused, only few people, high interaction

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Impression management in online dating

A

You want to promote yourself but also you want a small difference between your online and real self

  • users often create ideal self online and use these characteristics for self-growth
  • users use different identities linked to social context = identity as a social product
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Computer mediated communication - texting

A

No paralanguage: pitch, voice tone, volume, intonation…
- one tries to convey the same info by using grammar, abbreviations and emojis

WHY DO WE FAVOR TEXTING OVER CALLING?
control: impression, engagement, time, place

HYPERPERSONAL MODEL (Walther): CMC usage increases affection
Sender: selective self-representation
- no nonverbal leakage of info that doesn’t want to be shown
Channel: Discretionary engagement (self chosen time and mood to reply and editing of the message is possible)
Feedback: self fulfilling prophecy
Receiver: basic level of sympathy, over exaggeration of similarities, idealization
- fewer cues enhance perceived personal connection

SOCIAL PRESENCE THEORY: fewer cues = less connection experienced
Study: students of different units worked together over ST or LT and either had photos of the other students or not
- ST: photos had a positive effect on the perceived socio-emotional connection (social presence theory –> foto is the cue)
-LT: photo = negative effect (hyperpersonal model)

17
Q

Social robotics

A

Autonomous robot that interacts and communicates with humans and other autonomous physical agents by following social behaviors and rules attached to its role (a type of human robot interaction)

a) proximate interaction - close to robot (service)
b) remote interaction - far away (mars robot)

18
Q

Social robotics: the extent to which people anthropomorphize objects rests on

A
  1. SOCIALITY MOTIVATION: Human need to establish Social connections with others. If people are deprived of social connection, the humanize even non-human entities
  2. EFFECTANCE MOTIVATION: serves to satisfy need for mastery and control over given social environment
    - if our self-image of being competent is threatened, uncertainty arises. Particular likely if we encounter unknown agents (technical systems)
  3. ELICITED AGENT KNOWLEDGE: Participants more likely to humanize a system if it shares knowledge structure related to themselves
    - people use elicited agent knowledge to form a common ground with it by attributing human characteristics (in-group (- made in Germany)vs outgrip / gender /…)
19
Q

Human-robot interaction

A

Cognitive interaction technology: provide a deep assistance that enables intuitive human-machine interaction guided by principles of learning and memory

  • combines analysis of behavior and brain with synthetic approach
  • use robot platforms to test hypothesis about cognitive architecture

PROBLEM: still hard to build a system similar to human cognition and emotion

SOLUTION: needs to be tackled from different angles, (multidisciplinary approach)
- psychology needs to contribute with paradigms
Also new experimental ideas with clear control condition and clear manipulation of cause and effectt are needed

20
Q

Areas of application (human robot interaction)

A

TELEROBOTS: Human supervisory control of robots in performance of routine tasks
- limited series of actions automatically
Works by sensing environment and its own joint position and communicating that to the human operator
- set of eyes: easier for human to attend to what robot is doing

TELEOPERATION: remote control of space, airborne and undersea vehicles of non-routine tasks in hazardous environments - controlled by remote human

AUTOMATED VEHICLES WITH HUMAN PASSENGER - traffic is really difficult for AI to understand

HUMAN-ROBOT-SOCIAL INTERACTION

21
Q

why is it useful to model emotions?

A

emotions serve as a critical function in intelligent behavior

  • prime perception and response patterns
  • deficits in emotional processing = deficits in decision making
  • it is what motivates us to act (e.g. help)
22
Q

Computational models of emotion

A

address control and decision trade offs by directing cognitive resources toward the problem

  • how social emotions can facilitate interaction between human and machine
  • interested in modeling emotion in AI and robotics
23
Q

Bird example (computationally modeling human emotion)

A

Improvisational session with actors to record use of non verbal behavior
A dove flew in the room though the window:
1. surprise 2. fight or flight response 3. concern towards bird
Illustrates emotion function and re-prioritizing cognition
- action of actor served to adjust dynamic relation between actor and environment

24
Q

Challenges for emotional modeling

A
  1. the model must address how emotions arise and evolve over a range of eliciting conditions
  2. emotional responses can be rapid, or unfold over minutes or days

Emotion dynamics are linked with worlds dynamics and individual processes

25
Q

Theories of emotion

A

Many modern theories fall into 3 categories:

  1. Discrete theory of emotion: limited number of core emotions that are biologically determined and innate
    - expression shared across people and cultures
  2. Dimensional theories of emotion: emotions are point in a conceptual space - no specific biological basis
    These theories often use PAD theory with three dimensions:
    a) pleasure - measure of valence
    b) arousal - level of affective activation
    c) dominance - power and control
  3. Appraisal theory of emotion: emotions arise from a. process of comparing individual needs to external demands
    - person-environment relationship: characterized by appraisal variables (who caused it, is it even desirable…)
    PROBLEM: inference processes are needed before emotions can arise –> oppose automatic emotional reactions BUT irrelevant when viewed from a computational perspective: appraisal and inferences are distinct but operate on the same mental representation of person-environment relationship
26
Q

EMA model

A
  1. Appraisal derivation process - interprets person-environment relationship to derive appraisal variable
  2. Emotion derivation model: use appraisal value to form an emotional response

3 Behavioral response: coping strategies based on that emotion that subsequently manipulates person-environment

Person environment relation: model maintains an explicit representation of agent-environment relationship serving as input and output to various appraisal processes

  • referred to as causal interpretations: importance of causal reasoning on agents current knowledge of agent environment relationship (agents working memory)
  • set of beliefs, desires, plans, utilities and probabilities
Appraisal deprivation process: emotion eliciting events = appraisal variables: appraisal = fast, parallel and automatic. It maps features of causal interpretation into variables 
The appraisal frame maintains a continuously updated set of appraisal values associated with each proposition. Values include: 
a)	Relevance
b)	Desirability 
c)	Likelihood 
d)	Expectedness 
e)	Causal attribution 
f)	Controllability 
g)	Changeability

Emotion derivation: appraisal informs agents coping response but it is biased toward an overall mood state.

  • EMA maintains multiple appraisal frames, one for each proposition
  • individual appraisal frames build a mood

Affect consequences: A computational model of coping is included in the appraisal process. It determines the agents response to the appraised event and change person-environment relationship.
- coping strategies: control signals that enable or suppress cognitive processes that operate on causal interpretations

  1. Attention-related coping: Modulates agents attention to features of the environment, by altering their emotional state by altering WM
  2. Belief-related coping: Alterations to belief states in order to modulate these emotional states
  3. Desire-related coping: Regulates emotion by altering goal priorities
  4. Intention-related coping: Regulates emotions by altering intentions or by taking actions

Appraisal and coping operate within a loop, whereby situations may be appraised leading to emotions and coping responses that influence inferential processes altering person-environment relation

27
Q

EMA model: role in theory formation and testing

A

BENEFITS: powerful research tool, EMA makes appraisal simple, fast and general, maintains appraisal and coping shape without determining agents response, situation also plays a role.

Simon: emotions are an interrupt mechanism: appraisal and coping mediate response, so model is in accordance with Simon.

LIMITATIONS: more complex emotions aren’t modeled due to limited capacity of causal interpretations
- emotions are itself treated as epiphenomenal entities (nevertheless, its a powerful tool)

28
Q

Role of EMA in virtual human and HCI

A

Goal of EMA is to facilitate design and integration of multiple cognitive capabilities to design human like behavior. Use fo virtual humans allow to exploit it and respond to social emotions

29
Q

Validation of EMA

A

GOAL: create a model that can predict human emotion

Principles that guide empirical approaches to model human emotion:

  1. understanding relationship between emotions and unfolding situations - often mood and situation are investigated independently (not good)
  2. Linking intra- and interpersonal emotion: helps to coordinate individual and social behavior
  3. Model driven experimentation: create synergies between computational and psychological approaches

Man and Machine (5 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions