We are mining and modeling complex social systems using mobile devices.
We want to quantify the effect of social signals, social and physical
context, and content metadata on face-to-face interaction and sharing of
media content. By understanding individual's social behavior, we can
provide feedback for improved interaction and collaboration; auto-infer
privacy and sharing privileges; and suggest paths for widespread
distribution within the media network.
||Models of the diffusion of ideas, innovations, recommendations and media have been extensively studied in social science literature. We are defining new ways of studying social influence within the experimental context of media propagation, by using mobile phones to map patterns of behavior and social interaction and information flow.
Our mobile phone platform (iPhones and J2ME phones) mines activity and social interaction data for participants. In addition, users listen to music, rate and manage tracks, and forward it to other users through our (mobile) streaming indie music service.
Questions that we are trying to answer include - can we build more accurate models of diffusion and social influence using computational methods? Is it possible to look at behavioral patterns and identify 'social mavens' or 'opinion leaders'? Is it possible to auto-infer interest, privacy or sharing settings for users? By better understanding these questions, we hope to create more seamless and intuitive mobile media experiences.
||Taemie Kim, Agnes Chang
||We present the Meeting Mediator (MM), a real-time, personal, and portable system providing feedback to enhance group collaboration.
Social interactions are captured using Sociometric badges and are visualized on mobile phones to promote change in behavior.
In a study on brainstorming and problem-solving meetings, MM had a significant effect on overlapping speaking time and interactivity level without distracting the subjects.
Our system encourages effective group dynamics that may lead to higher performance and satisfaction.
We envision MM to be deployed in real-world organizations to improve interactions across various group collaboration contexts.
||Anmol Madan (Alumni Contributors: Ron Caneel, Will Stollzman, Jon Gips)
||MNon-linguistic social signals (e.g. 'tone of voice') are often as important as linguistic content in predicting behavioral and social outcomes. In this research project, we propose automated measures of social signaling based on tone and prosody in voice.
We have evaluated the performance of these automated measures in predicting interest in conversation, outcomes of business negotiations, dating encounters, contact center interactions, and social interactions. We have also devised real-time feedback technologies on mobile phones to allow users to improve their real-time interactions.
Pentland A."Honest Signals" (in publication).
Madan A. and Pentland A."Mob.Media: A Mobile Phone Platform for Computational Social Science" (in submission)
T. Kim, A. Chang, L. Holland, A. Pentland "Meeting Mediator: Enhancing Group Collaboration and Leadership with
Sociometric Feedback" . In Proceedings of CSCW, San Diego, CA, November 2008.
T. Kim, A. Chang, L. Holland, A. Pentland "Meeting Mediator: Enhancing Group Collaboration with
Sociometric Feedback".CHI Extended Abstracts, Florence, Italy, April 2008.
T. Kim, A. Chang, A. Pentland, "Enhancing Organizational Communication using Sociometric
Badges" . IEEE 11th International Symposium on Wearable Computing (Doctoral Colloquium
Proceedings), Boston MA, October 2007.
Madan A., and Pentland A.,"VibeFones: Socially Aware Mobile Phones". International Symposium of Wearable Computers, ISWC 2006 .
Madan A., "Thin Slices of Interest". Master's Thesis in the Media Laboratory (Advisor Dr. Alex Pentland), June 2005.
Caneel R. "Social Signaling in Decision Making". Master's Thesis in the Media Laboratory, June 2005.
Madan A., Caneel R. and Pentland A.,"Voices of Attraction". Proceedings of Augmented Cognition, HCI 2005, Las Vegas.
Pentland A.,"Social Dynamics: Signals and Behavior". MIT Media Lab Technote 579.