affect, appeal, and sentiment as factors influencing interaction - - PowerPoint PPT Presentation

affect appeal and sentiment as factors influencing
SMART_READER_LITE
LIVE PREVIEW

affect, appeal, and sentiment as factors influencing interaction - - PowerPoint PPT Presentation

affect, appeal, and sentiment as factors influencing interaction with multimedia information jussi karlgren september 29, 2009, korfu results from the chorus coordination action plenty of recent research on affect in users by computational


slide-1
SLIDE 1

affect, appeal, and sentiment as factors influencing interaction with multimedia information

jussi karlgren september 29, 2009, korfu

slide-2
SLIDE 2

results from the chorus coordination action

slide-3
SLIDE 3

plenty of recent research on affect in users by computational research

slide-4
SLIDE 4

starting points

1 information items (may) carry explicit or implicit sentimental

loading

2 people are in continuously changing affective states 3 activities people engage in modify and are modified by the

affective state of the user.

slide-5
SLIDE 5

starting points

1 information items (may) carry explicit or implicit sentimental

loading

2 people are in continuously changing affective states 3 activities people engage in modify and are modified by the

affective state of the user.

slide-6
SLIDE 6

several reasons to study affect, appeal, and sentiment

1 better interactive systems in general, as systems move from

  • ffice to general use

2 next generation of computer mediated human-human

communication

3 dynamic narratives, for entertainment and education 4 understanding people better 5 more concrete task: information access

(that’s us)

slide-7
SLIDE 7

several reasons to study affect, appeal, and sentiment

1 better interactive systems in general, as systems move from

  • ffice to general use

2 next generation of computer mediated human-human

communication

3 dynamic narratives, for entertainment and education 4 understanding people better 5 more concrete task: information access

(that’s us)

slide-8
SLIDE 8

several reasons to study affect, appeal, and sentiment

1 better interactive systems in general, as systems move from

  • ffice to general use

2 next generation of computer mediated human-human

communication

3 dynamic narratives, for entertainment and education 4 understanding people better 5 more concrete task: information access

(that’s us)

slide-9
SLIDE 9

several reasons to study affect, appeal, and sentiment

1 better interactive systems in general, as systems move from

  • ffice to general use

2 next generation of computer mediated human-human

communication

3 dynamic narratives, for entertainment and education 4 understanding people better 5 more concrete task: information access

(that’s us)

slide-10
SLIDE 10

several reasons to study affect, appeal, and sentiment

1 better interactive systems in general, as systems move from

  • ffice to general use

2 next generation of computer mediated human-human

communication

3 dynamic narratives, for entertainment and education 4 understanding people better 5 more concrete task: information access

(that’s us)

slide-11
SLIDE 11

several reasons to study affect, appeal, and sentiment

1 better interactive systems in general, as systems move from

  • ffice to general use

2 next generation of computer mediated human-human

communication

3 dynamic narratives, for entertainment and education 4 understanding people better 5 more concrete task: information access

(that’s us)

slide-12
SLIDE 12

workshop on affect, appeal, and sentiment as factors influencing interaction with multimedia information may 28, 2009 brussels

slide-13
SLIDE 13

some discussion points

1 useful use cases 2 appropriate models 3 enhanced target notion for retrieval evaluation 4 link between usage and data objects

slide-14
SLIDE 14

some discussion points

1 useful use cases 2 appropriate models 3 enhanced target notion for retrieval evaluation 4 link between usage and data objects

slide-15
SLIDE 15

some discussion points

1 useful use cases 2 appropriate models 3 enhanced target notion for retrieval evaluation 4 link between usage and data objects

slide-16
SLIDE 16

some discussion points

1 useful use cases 2 appropriate models 3 enhanced target notion for retrieval evaluation 4 link between usage and data objects

slide-17
SLIDE 17

(some) use case related factors motivation task view broken by multimedia situation lean-back, not lean-forward success satisfaction not optimisation effort short-coding, not specification

slide-18
SLIDE 18

(some) use case related factors motivation task view broken by multimedia situation lean-back, not lean-forward success satisfaction not optimisation effort short-coding, not specification

slide-19
SLIDE 19

(some) use case related factors motivation task view broken by multimedia situation lean-back, not lean-forward success satisfaction not optimisation effort short-coding, not specification

slide-20
SLIDE 20

(some) use case related factors motivation task view broken by multimedia situation lean-back, not lean-forward success satisfaction not optimisation effort short-coding, not specification

slide-21
SLIDE 21

(some) use case related factors motivation task view broken by multimedia situation lean-back, not lean-forward success satisfaction not optimisation effort short-coding, not specification

slide-22
SLIDE 22

initial take on models of affect

1 categorial models with palettes of emotions 2 dimensional representations with emotions projected in low

dimensional space not the last word - more dynamic models likely

slide-23
SLIDE 23

initial take on models of affect

1 categorial models with palettes of emotions 2 dimensional representations with emotions projected in low

dimensional space not the last word - more dynamic models likely

slide-24
SLIDE 24

initial take on models of affect

1 categorial models with palettes of emotions 2 dimensional representations with emotions projected in low

dimensional space not the last word - more dynamic models likely

slide-25
SLIDE 25

initial take on models of affect

1 categorial models with palettes of emotions 2 dimensional representations with emotions projected in low

dimensional space not the last word - more dynamic models likely

slide-26
SLIDE 26

tentative goal for information access research reformed target notion for evaluation: introduce appeal to complement relevance include factors related to user in evaluation, not only relations document and specification

slide-27
SLIDE 27

tentative goal for information access research reformed target notion for evaluation: introduce appeal to complement relevance include factors related to user in evaluation, not only relations document and specification

slide-28
SLIDE 28

tentative goal for information access research reformed target notion for evaluation: introduce appeal to complement relevance include factors related to user in evaluation, not only relations document and specification

slide-29
SLIDE 29

tenuous link between usage and data objects recall:

1 information items (may) carry explicit or implicit sentimental loading 2 people are in continuously changing affective states 3 activities people engage in modify and are modified by the affective

state of the user

what is the link? what outcome variables can we use? what ground truth might we have? IAPS, but for txt?

slide-30
SLIDE 30

tenuous link between usage and data objects recall:

1 information items (may) carry explicit or implicit sentimental loading 2 people are in continuously changing affective states 3 activities people engage in modify and are modified by the affective

state of the user

what is the link? what outcome variables can we use? what ground truth might we have? IAPS, but for txt?

slide-31
SLIDE 31

tenuous link between usage and data objects recall:

1 information items (may) carry explicit or implicit sentimental loading 2 people are in continuously changing affective states 3 activities people engage in modify and are modified by the affective

state of the user

what is the link? what outcome variables can we use? what ground truth might we have? IAPS, but for txt?

slide-32
SLIDE 32

tenuous link between usage and data objects recall:

1 information items (may) carry explicit or implicit sentimental loading 2 people are in continuously changing affective states 3 activities people engage in modify and are modified by the affective

state of the user

what is the link? what outcome variables can we use? what ground truth might we have? IAPS, but for txt?

slide-33
SLIDE 33

first steps being taken now.