University of Southern California
Research Group

Workshop on Populations & Crowds: Dynamics, Disruptions and their Computational Models

Dates: September 6-7, 2012

Location: RTH Boardroom (RTH 526)


Milind Tambe,
Professor of Computer Science and Industrial & Systems Engineering Departments,
University of Southern California

P. Jeffrey Brantingham,
Professor of Anthropology, UCLA

Sponsored by the Army Research Office



Populations and crowds appear to share many features in common. Both are aggregates of individuals that display group-level organization in space and time. Both are systems through which cultural information and beliefs, as well as genes and diseases, may propagate. And both are capable of turning from benign to hostile. Despite these many similarities, populations and crowds are often treated separately in theoretical and empirical studies. The different temporal and spatial scales at which people conceive of populations and crowds are at least partially responsible for this conceptual separation; the former evolve on time scales of years or more and occupy sometimes vast spatial regions, while the latter may evolve on time scales of minutes to hours and are generally spatially compact phenomena. The purpose of this workshop is to bring together leading researchers from both computational and social sciences with expertise in domains relevant to modeling of populations and crowds. Our interest is in understanding how populations and crowds form, how they are organized, how beliefs propagate through them, what underlies tipping points their behavior and, ultimately, what can be done to block contagion of hostile behavior in both population and crowd contexts. The workshop will be held at the University of Southern California on September 6-7, 2012.
Image Image Image Image

Motivation & Goals

As recent events in across the greater Middle East and North Africa make clear, both populations and crowds can wield tremendous physical, social, economic and ideological power. Crowds may assemble quickly and dramatically alter the course of history. The protests in Sidi Buazid, Tunisia, Cairo’s Tahrir Square, and Sana, Yemen, each grew from a few hundred individuals to tens of thousands within days of the first signs of unrest. Some of the power of the crowd stems simply from the physical mass of people involved, which may give the crowd the ability to withstand and even overwhelm all but the most extreme kinetic crowd control measures. However, crowds are also powerful in their ability to spread information and rapidly alter their collective behavior. Crowds can transition from loosely to tightly organized and benign to hostile in a matter of moments. They can also disperse as rapidly as they form, making their power also temporary. The political, economic and social stability of an area often hinges on the alignment and actions of its resident populations. Single events, both local and remote, can lead populations to align, or realign themselves and take action. The gradual spread of information in the form of secular or sacred values across a population may lead to similar realignments over the long run. Ultimately, populations are both repositories for an effective mechanisms for transmitting cultural (and biological) information. In the absence of strong secular institutions, populations may organize along sectarian and cultural lines—witness the dramatic transformation of Baghdad from integrated Sunni-Shia neighborhoods to a largely segregated city from 2003-2006. They can also collapse under internal and external pressures.
We seek to understand the organization and dynamics of populations and crowds in both qualitative and quantitative terms. Yet, despite their many apparent similarities, populations and crowds are often treated separately in theoretical and empirical studies. This may in part relate to the different spatial and temporal scales at which we perceive populations and crowds. Consequently, researchers studying population organization and dynamics may not be fully aware of how their work relates to those studying crowds, and vice versa. Potentially important theoretical and methodological advances bearing on the control and disruption of hostile populations and crowds may be lost in the gap between domains.

To bridge this gap, we propose to hold a cross-disciplinary workshop at the University of Southern California on 6-7 September, 2012 concentrating on organization, dynamics and disruption of populations and crowds. The purpose of this workshop is to bring together leading researchers from both computational and social sciences with expertise in domains relevant to modeling of populations and crowds. The goal will be to encourage the development of synergistic theory and methods that go beyond what either domain currently deploys.
In this regard, the workshop will focus on a range of questions including but not limited to:

1. Are typological distinctions between populations and crowds useful? Or, do populations and crowds display regular scaling relationships with respect to one another in time and space?
2. Do the differing temporal and spatial scales of populations and crowds create divergent opportunities for positive and negative assortment, or other organizational differences?
3. Do beliefs propagate in the same ways through populations and crowds and, if not, what structural or dynamic properties of populations and crowds explain these differences? Again, what types of computational techniques and formalisms could be used to model such belief propagation?
4. How does the injection of social-media into crowd & population systems, as opposed to face-to-face interactions, impact crowd & population dynamics? What types of actions can help influence competitive contagion or in blocking such contagion?
5. What types of computational techniques and formalisms can be used to characterize and model the dynamic structural changes exhibited by populations and crowds?
6. What computational techniques are useful for modeling and analysis of competitive contagion of beliefs and ways to block belief contagion in both crowds and populations? What role does analysis of social networks play in such modeling?
7. What are some research challenges in constructing such computational models? How do we validate these computational models?


The 2-day workshop will be held in 6-7 September 2012 at the University of Southern California (USC) in Los Angeles. Staff support will be provided by USC. There will be 15 to 20 academic participants.

In addition, there will be about 10 participants from the US military, government organizations, and selected defense organizations. We will host a moderated panel discussion on policing crowds with law enforcement members representing US, UK and Indian Police Forces.

The anticipated list of speakers from academia and law enforcement agencies in Los Angeles includes:
1. John O'Brien, Associate Dean, USC Viterbi School of Engineering
2. Purush Iyer, ARO
3. Chjan Lim, Mathematical Sciences, RPI Presentation
4. Jenna Bednar, Political Science, University of Michigan Presentation
5. Rebecca Goolsby, Office of Naval Research
6. Rajiv Maheswaran, Yu-han Chang, Information Sciences Institute, USC Presentation
7. Nicholas Weller, Mathew McCubbins, USC Presentation
8. Mubarak Shah, Computer Science, Univerity of Central Florida Presentation
9. Mark Turner Presentation
10. Shashi Shekar, Computer Science, University of Minnesota
11. Maj Benjamin Hung Presentation
12. David Kempe, Milind Tambe, USC Computer Science Presentation by Prof. Tambe
13. Gal Kaminka, Multidisciplinary Brain Research Institute, Bar Ilan University, Israel Presentation
14. Panel Discussion:
  • Bob Green, Los Angeles Police Department
  • John Sullivan, Los Angeles County Sherriff's Department
  • Arvind Verma, Criminology, Indiana University
15. Andrea Bertozzi, Mathematics, UCLA Presentation Clip1 Clip2 Clip3 Clip4 Clip5
16. Anders Johansson, Systems Engineering, University of Bristol
17. Jeff Brantingham, UCLA Anthropology Presentation

In addition, several key researchers from federal government agencies have also agreed to participate in the workshop.
These individuals include:
18. Chris Arney, USMA
19. Kate Coronges, USMA
20. Hillary Fletcher, USMA
21. Jeff Johnson, Army Research Office
22. Brian Lande, DARPA
23. Anantharam Swamy, Army Research Laboratory
24. Rand Waltzman, DARPA
25. Bruce West, Army Research Office
26. Jimmie Jaye Wells, 75th training division, mission command
27. Fredrick Clark, ARO