Human-Computer dialogue Developing a Dialogue Editor to Script - - PowerPoint PPT Presentation

human computer dialogue
SMART_READER_LITE
LIVE PREVIEW

Human-Computer dialogue Developing a Dialogue Editor to Script - - PowerPoint PPT Presentation

Human-Computer dialogue Developing a Dialogue Editor to Script Interaction between Virtual Characters and Social Phobic Patients !


slide-1
SLIDE 1
  • Human-Computer dialogue

Developing a Dialogue Editor to Script Interaction between Virtual Characters and Social Phobic Patients

! "# #$"%&$

slide-2
SLIDE 2
  • Human-Computer Dialogue

Overview

  • VRET
  • Social phobia
  • Dialogue
  • Creation
  • Editor
  • Evaluation
  • Pilot study
  • Setup
  • Results
  • Experiment
  • Preliminary results
slide-3
SLIDE 3
  • Human-Computer Dialogue

1.

Virtual Reality Exposure Therapy

VRET

slide-4
SLIDE 4

' Human-Computer Dialogue

VRET

( % )* ( *+,%

  • -

(*+,

  • .. /0
  • 102
  • 02
  • "/
  • /"2

Phobias

"3

slide-5
SLIDE 5

4 Human-Computer Dialogue

0$05667773023$

slide-6
SLIDE 6

8 Human-Computer Dialogue

VRET

( /"2

  • 9/00
  • )$06

( %

  • $

( ,:*+,7

  • "01
  • $01$
  • ;$0

Social Phobia

<3 ==' $3 ===

slide-7
SLIDE 7

> Human-Computer Dialogue

slide-8
SLIDE 8
  • Human-Computer Dialogue
slide-9
SLIDE 9

= Human-Computer Dialogue

VRET

( ./0

  • .0
  • /0
  • :
  • /$00
  • .?@0?0@
  • <7?$0@

Social Phobia

slide-10
SLIDE 10
  • Human-Computer Dialogue

2.

Dialogue

slide-11
SLIDE 11
  • Human-Computer Dialogue

Dialogue

  • Structured and controlled dialogue
  • Question -> Response structure
  • Neutral topic
  • Guiding Questions
  • Response groups
  • Keyword
  • Yes/No
  • Attitude
  • Unknown
  • Length
  • General
  • Question

Creation

slide-12
SLIDE 12
  • Human-Computer Dialogue

Dialogue

Editor

(

  • Computer -> User -> Computer
  • Question -> Response -> Computer Response
  • Complexity grows quickly
  • Special tool needed
slide-13
SLIDE 13
  • Human-Computer Dialogue

Dialogue

( ( ,

  • ,
  • /002$
  • -$
  • $
  • /
  • /!;2

( A0

Editor3

slide-14
SLIDE 14

' Human-Computer Dialogue

slide-15
SLIDE 15

4 Human-Computer Dialogue

slide-16
SLIDE 16

8 Human-Computer Dialogue

slide-17
SLIDE 17

> Human-Computer Dialogue

slide-18
SLIDE 18
  • Human-Computer Dialogue

Dialogue

  • Iterative design dialogue
  • MSN Chatbot
  • 4 topics selected
  • Keywords added

Evaluation

  • !

"!! #$%& &

  • '($!

'($! ($! %)%! * '($! &! #+,- !!

$

= 3

  • '
  • 3

?3=@

.

  • 34
  • 38

?3@

  • 3=
  • 4
  • 3

?3@

"

  • 38
  • 8

8 4 3 ?3@

slide-19
SLIDE 19

= Human-Computer Dialogue

3.

Pilot Study

slide-20
SLIDE 20
  • Human-Computer Dialogue

Pilot Study

  • 4 Participants
  • 3D world projected on screen
  • Participants alone in the room
  • Audio interaction
  • 4 Testing conditions
  • Speech detection
  • Speech recognition (limited keywords)
  • Speech recognition (keywords)
  • Human controlled

Setup (Movie)

slide-21
SLIDE 21
  • Human-Computer Dialogue

Pilot Study

  • Used same keywords
  • Spoke longer and more elaborate
  • Responses on content increased realism
  • Waiting time before response decrease realism

Results

slide-22
SLIDE 22
  • Human-Computer Dialogue

3.

Experiment

slide-23
SLIDE 23
  • Human-Computer Dialogue

Experiment

  • 24 participants
  • Dutch, Adults
  • Participant remarks
  • Short avatar introduction of them selves
  • To intensive staring
  • Responses on content increase presence (reported)
  • Delay in response decreased presence (reported)

Setup / remarks

slide-24
SLIDE 24

' Human-Computer Dialogue

Experiment

  • Dialog interaction speed
  • F(3,21) = 11.49

p = 0.001

Preliminary results

slide-25
SLIDE 25

4 Human-Computer Dialogue

Experiment

  • Interaction Feeling
  • F(3,21) = 11.878

p = 0.001

Preliminary results

slide-26
SLIDE 26

8 Human-Computer Dialogue

Experiment

  • Interaction Reality
  • F(3,21) = 3.921

p = 0.023

  • Speech Detector

/ Human Control F(1,23) = 0.538

p = 0.471

Preliminary results

slide-27
SLIDE 27

> Human-Computer Dialogue

Experiment

  • Dialog Flow Interruptions
  • F(3,21) = 3.208

p = 0.044 Therapist liked the interruptions

Preliminary results

slide-28
SLIDE 28
  • Human-Computer Dialogue

Future improvements

  • Reponses on long answers
  • Allow questions
  • Decrease response times
  • Analyze when speaking
  • Faster protocols
  • Sharper timings
  • More natural avatar movements
  • Anxiety increasing computer responses
slide-29
SLIDE 29

= Human-Computer Dialogue

Literature

( " 33 B+CC 1313?@31$:0 0:00251$3D%0 ,:0$" =?@ 483 ( < +33 - <313 E /3F3 3%3 -3 ,$ /3 3?=='@3;$-"$)+" / +$$2/31G " 4?@ = ( $ /3G3 B%7 33?===@3:5)$0 0300 ''3 ( < ,3 % /3 ; "3 + /3 ; .3 $ )3 3?4@3 *02002510$ 320B% ?@ >83 ( %&$ #3"3 - 33 313"3G3 B* 3?@3*:0 002510&73" )7&0 ) 43

slide-30
SLIDE 30
  • Human-Computer dialogue

Developing a Dialogue Editor to Script Interaction between Virtual Characters and Social Phobic Patients

! "# #$"%&$