Title: Persistent Social Structures
Speaker: Tanya Berger-Wolf, DIMACS postdoc
Date: July 5, 2005 12:30 - 2:00 pm
Location: DIMACS Center, CoRE Bldg, Room 431, Rutgers University, Busch Campus, Piscataway, NJ
Abstract:
Humans, zebras, geese, ants, bacteria, all form social groups. Taken as a snapshot any population of social animals will have several groups in it. However, a snapshot taken at a different time may show totally different groups. Social associations change from minute to minute, hour to hour, day to day. Some groups tend to change more, while others, less. Some just loose or gain a few members at a time while others break up into many smaller groups and reassemble again. Other still just reappear every now and then. Which social groups are important, significant, stable? What does it mean for a group to persist over time? And over what time? We try to answer these questions algorithmically, modeling populations as dynamic graphs. We give a formal general definition of persistence of a social group and try to analyze what it means in various contexts and what algorithms are appropriate for various applications. Our main application right now is the social structure of zebras at the Mpala Conservancy in Kenya, A frica, and onagers in India. This project is in collaboration with Dan Rubenstein's group from Princeton, Ared Saia from University of New Mexico and S. Muthukrishnan from Rutgers University.