Post on 11-Jan-2016
transcript
© 2006 MIT Media Lab
Social Network Technology to Evaluate and Facilitate
Collaboration
MIT Media LabHuman Dynamics GroupProf. Alex (Sandy) Pentland
Daniel Olguin OlguinMichael Sung
NIH Roadmap Interdisciplinary Methodology and Technology SummitNorth Bethesda, MD August 21-22, 2006
© 2006 MIT Media Lab
Human Dynamics Research
GroupMEDIA
Learning Humans
LiveNet
Reality Mining
Sensible Organizations
© 2006 MIT Media Lab
Underlying Framework Social signals
– From speech: engagement, emphasis, mirroring, activity
– From body gesture: motion, energy, activity
We have been able to identify:– Central connectors, boundary
spanners, information brokers and peripheral people in a social network
– The boss in an organization– The leader of a team– The outcome of negotiations– The degree of persuasiveness in
speech– Group affiliations
Automatically captured group dynamics
© 2006 MIT Media Lab
Wearable Computing
Electronic Badges
Body Sensor Networks
Human Activity Recognition
Healthcare ApplicationsMIThrill
Body motion
Face-to face interactions
© 2006 MIT Media Lab
Social Motion
Conferences and Career Fairs
Identifying team leaders and experts
Affiliation and Social Relationship Inference
Automatic Real-timeInterest Measurement
© 2006 MIT Media Lab
LifeWear
Human Activity Recognition Using Wearable Sensors– PDA– Camera– Microphones– Accelerometers
Automatic Multimedia Collection of Interesting Moments
© 2006 MIT Media Lab
Healthcare Applications
DiaBetNet– Wearable computer
for diabetic children Wearable Monitor for
Parkinson Disease Treatment
LiveNet DiaBetNet: Interactive game to monitor blood
glucose levels and make predictions
© 2006 MIT Media Lab
GroupMEDIAAdding Context Awareness to Mobile
Devices
Modeling User Behavior
Classification Accuracy: 80-
90%The “Jerk-o-Meter”
Speed Dating
© 2006 MIT Media Lab
Reality MiningEigenbehaviors: Identifying structure in routine
Proximity Sensing:Bluetooth + Cell tower ID
Social Serendipity
© 2006 MIT Media Lab
Sensible Organizations
Understanding Organizational
Dynamics
Efficiency
Creativity
Productivity
Innovation
Capturing everyday social signals in real
organizations to improve managerial practicesUsing social sensors
technology to measure:
Combining social,
physical, and digital
information
© 2006 MIT Media Lab
Social Sensors Technology Extended mobile phones:
– Bluetooth-enabled smart-phones– Wearable electronic badges with social sensors
Real-time speech feature analysis Context awareness, user localization and proximity
sensing Activity recognition Push to talk system with voice-controlled interface
Mobile phones are socially accepted wearable computers
© 2006 MIT Media Lab
Wearable Communicator Badge
© 2006 MIT Media Lab
Technological Challenges
User acceptance– Small and comfortable to wear
Hardware design, development, and support– Prototyping, manufacturing, and
deployment Real time data collection and
processing Large-scale user studies
© 2006 MIT Media Lab
Methodological Challenges Relate social measurements to productivity,
efficiency, creativity and innovation– Develop new metrics to achieve quantitative
measurements– Evaluate qualitative data: consumer satisfaction
Perform dynamic social network analysis Capture individual and group dynamics in
locally and geographically distributed teams New management methodologies based on
social sensors
© 2006 MIT Media Lab
The LiveNet System Distributed modular framework Commodity PDA/cell phone hardware Variety of custom/commercial sensors Real-time data streaming Resource allocation/discovery Local processing for context
classification Rapid application prototyping
LiveNet: a flexible mobile platform that is at the same time a long-term health monitor, context-aware agent, multi-modal feedback interface for proactive healthcare applications
© 2006 MIT Media Lab
Non-invasive Sensing Movement
– spectral features, energy, orientation Voice Features
– energy, pitch, entropy, voicing dynamics
Temperature/heat flux– Metabolic activity, environmental
cues Heart rate
– IBI, HRV measures, spectral ratios Skin conductance
– slope analysis, peak detection Behavioral
– Location, sleep/activity patterns, socialization dynamics
BioSense Board
Bluetooth Location Beacon
© 2006 MIT Media Lab
MIT PokerMetrics Stress Study
LiveNet PokerMetrics Setup Real-time Physiology (Stressful vs Non-Stressful)
© 2006 MIT Media Lab
U.S. Army Soldier Physiology Monitoring
LiveNet ARIEM System Shivering Core Temperature Regimes
© 2006 MIT Media Lab
MGH Depression and ECT Treatment Study
LiveNet Depression Rig
Subjective emotion ratingsClinical Outcomes
Physiology correlations (1 day)Emotion rating correlations
© 2006 MIT Media Lab
ThanksFor more information visit:
http://hd.media.mit.eduor e-mail us at:
dolguin@media.mit.edu