Effect of Shared-attention for Human-Robot Interaction

26
Effect of Shared-attention for Human-Robot Interaction Junji Yamato [email protected] NTT Communication Science Labs., NTT Corp. Japan Kazuhiko Shinozawa, Futoshi Naya ATR Intelligent Robot and Communication Labs.

description

Effect of Shared-attention for Human-Robot Interaction. Junji Yamato [email protected] NTT Communication Science Labs., NTT Corp. Japan Kazuhiko Shinozawa, Futoshi Naya ATR Intelligent Robot and Communication Labs. Aim. To build Social Robot/Agent Sub goal To establish Evaluation methods - PowerPoint PPT Presentation

Transcript of Effect of Shared-attention for Human-Robot Interaction

Page 1: Effect of Shared-attention for Human-Robot Interaction

Effect of Shared-attention for Human-Robot Interaction

Junji Yamato [email protected]  

NTT Communication Science Labs., NTT Corp. JapanKazuhiko Shinozawa, Futoshi NayaATR Intelligent Robot and Communication Labs.

Page 2: Effect of Shared-attention for Human-Robot Interaction

Aim

To build Social Robot/Agent

Sub goalTo establish

Evaluation methodsDesign guidelines

for communication of human-robot/agent

Page 3: Effect of Shared-attention for Human-Robot Interaction

• To measure the influence of Agent/Robot on users

• Acceptance ratio of agent/robot recommendation

Method

Page 4: Effect of Shared-attention for Human-Robot Interaction

Color name selection taskBlue or Green? Cobalt green or emerald green?Skin color or KARE-IRO?SUMIRE-IRO or AYAME-IRO?--------

Total:30 questions.(from color name text book)

• No “correct” answer• Easy to be influenced

Page 5: Effect of Shared-attention for Human-Robot Interaction

Four experiments1. Compared agent and robot2. Compared agent and robot in physical

world3. Measured the effect of eye contact4. Measured the effect of shared-attention

Yamato, J., Shinozawa, K., Brooks, R., and Naya, F. Human-Robot Dynamic Social Interaction. NTT Technical Review 1, 6(2003), 37-43.

Detailed description of Experiment 1 and 2

Available on-linehttp://www.ntt.co.jp/tr/Back number -> Sep. 2003

Shinozawa, K., Naya, F., Yamato, J., and Kogure, K. Differences in Effect of Robot and Screen Agent Recommendations on Human Decision-Making , IJHCS (to appear)

Experiment 1, 2, and description of K4(robot)

Page 6: Effect of Shared-attention for Human-Robot Interaction

Experiment 1 :Compare Agent and Robot

• Conditions: 30 questions, 30 subjects in each group‐ Same question sequences, same voice, similar gesture• Measurement: acceptance ratio, questionnaire

Agent RobotAgent Robot

Page 7: Effect of Shared-attention for Human-Robot Interaction

Experiment 1: Robot

QuickTime˛ Ç∆MPEG-4 ÉrÉfÉI êLí£ÉvÉçÉOÉâÉÄ

ǙDZÇÃÉsÉNÉ`ÉÉÇ å©ÇÈÇΩÇflÇ…ÇÕïKóvÇ≈Ç∑ÅB

Page 8: Effect of Shared-attention for Human-Robot Interaction

Experiment 1: Result

• Acceptance : agent > robot (p<.01)

• Familiarity : independent

Page 9: Effect of Shared-attention for Human-Robot Interaction

Initial expectationRobot has more influence because it lives in 3D world,same as subjects.

agentrobot

×

GapGap Gap

Page 10: Effect of Shared-attention for Human-Robot Interaction

Experiment 2: Compare in physical world

Color plate

Button box

Button box

• No recommendation (30 subjects)• Recommendation by robot ( 31 subjects )• Recommendation by agent (30 subjects)

Page 11: Effect of Shared-attention for Human-Robot Interaction

Experiments

Page 12: Effect of Shared-attention for Human-Robot Interaction

Experiment 2: Result

• selection ratio : robot > agent ( p < 0.05)       robot>> no recommendation ( p < 0.01)

Page 13: Effect of Shared-attention for Human-Robot Interaction

Experiment 1 and 2: Results

Media world Physical worldagentrobot

Consistency matters.

× ○

×

Embodiment and communication

Page 14: Effect of Shared-attention for Human-Robot Interaction

Why robot is better?

• Easy to detect gaze– Eye contact– Shared attention/joint attention

Measure the effect of eye contact and shared-attention

Page 15: Effect of Shared-attention for Human-Robot Interaction

• Eye contact was established by face tracking

• Eye contact time: period that subject looked at robot and robot looked at subject

• Eye contact time and selection ratio?• Two groups (14 subjects each)

– Eye contact, and NO eye contact

Experiment 3: Effect of eye contact (mutual gaze)

Page 16: Effect of Shared-attention for Human-Robot Interaction

Robots

Page 17: Effect of Shared-attention for Human-Robot Interaction

• Higher selection ratio for eye contact group• K4: No E.C. < E.C. (p=0.012)• Rabbit: No E.C. < E.C. (p=0.003)

Selection ratio

Page 18: Effect of Shared-attention for Human-Robot Interaction

• Shared attention:– Period that robot looks at an object and

subject looks at the same object. (color plate, button box)

• SA time and selection ratio– Is there correlation?

Experiment 4: Effect of shared-attention

Page 19: Effect of Shared-attention for Human-Robot Interaction

• Robot looks at color plate and button box by prepared program

• Eye contact established by face tracking

Establishing shared-attention

QuickTime˛ Ç∆ÉtÉHÉg - JPEG êLí£ÉvÉçÉOÉâÉÄǙDZÇÃÉsÉNÉ`ÉÉÇ å©ÇÈÇΩÇflÇ…ÇÕïKóvÇ≈Ç∑ÅB

QuickTime˛ Ç∆ÉtÉHÉg - JPEG êLí£ÉvÉçÉOÉâÉÄǙDZÇÃÉsÉNÉ`ÉÉÇ å©ÇÈÇΩÇflÇ…ÇÕïKóvÇ≈Ç∑ÅB

QuickTime˛ Ç∆ÉtÉHÉg - JPEG êLí£ÉvÉçÉOÉâÉÄǙDZÇÃÉsÉNÉ`ÉÉÇ å©ÇÈÇΩÇflÇ…ÇÕïKóvÇ≈Ç∑ÅB

Example: video

Page 20: Effect of Shared-attention for Human-Robot Interaction

• 28 subjects• SA time = 51.7 sec (total for 30 questions)

– (Longer than in Experiment 3 )• Selection ratio. Average: 0.57 S.D.= 0.14

• Some subjects were positive, and others were not. Clear contrast, from the questionnaire.

   Example: Robot is prompting wrong choice. I feel the robot forced me to select his recommendation (negative).

Experimental conditions

Page 21: Effect of Shared-attention for Human-Robot Interaction

• No correlation

SA time and selection ratio

Shared-Attention time (count) 50count=1sec.

Selection ratio

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 500 1000 1500 2000 2500 3000 3500 4000

ånóÒ1

Page 22: Effect of Shared-attention for Human-Robot Interaction

• Ego-gram based on transactional analysis

• Measure three ego-states by questionnaire– CP, NP (critical parent, nurturing

parent)– A (adult)– FC, AC (free child, adapted child)

• TEG (Tokyo Univ. Egogram ) is common in Japan

Clustering subjects by TEG(Ego-gram)

Page 23: Effect of Shared-attention for Human-Robot Interaction

• Strong correlation in SA time and acceptance ratio for high AC (Adapted Child) group

High/Low TEG measurement and SA time.

CP NP A FC AC

Threshold (threshold) 10 17 12 14 12

Selection ratio High group 0.548717949 0.564444444 0.564285714 0.617948718 0.597777778

Low group 0.586666667 0.574358974 0.573809524 0.526666667 0.535897436

SAtime2553.461538 2581.933333 2575.285714 2587.615385 2510.866667

2612.266667 2588.461538 2594.642857 2582.666667 2670.461538

Correlation fo High -0.046541377 -0.035889649 -0.03145821 0.136152737 0.405908084

SAtime and sel ratio Low 0.054648939 0.10996751 0.055147415 -0.078258186 -0.461514856

Page 24: Effect of Shared-attention for Human-Robot Interaction

• Positive correlation(Speaman’s r=0.51,p=0.051).

SA time and selection ratio (high AC & low CP group)

Page 25: Effect of Shared-attention for Human-Robot Interaction

• High-SA group = high selection ratio (p<0.05)

SA time and selection ratio

( high AC group)

Page 26: Effect of Shared-attention for Human-Robot Interaction

• High AC subject (obedient type) showed positive correlation between SA time and selection ratio.

• No significant difference between SA time itself and selection ratios for high AC and low AC groups

• Eye contact and shared-attention promote close communication. Some people like such intimate relation, but others don’t. It depends on the character.

• SA is effective. Even SA was not “actually” realized. We do not need to develop image understanding technology; we just have to fake it.

Result and Discussion