DIMACS/CCICADA Student Workshop on Where the Mathematical and Computational Sciences Meet Society

April 8, 2011
DIMACS Center, CoRE Building, Rutgers University

Student Organizers:
Ed Chien, Rutgers University, (Math)
Bobby DeMarco, Rutgers University, (Math)
Brad Greening, Rutgers University, (Ecology)
Emilie Hogan, Rutgers University, (Math)
Scott Kulp, Rutgers University, (CS)
Matthew Oster, Rutgers University, (Operations Research)
Brian Thompson, Rutgers University, (CS)
Faculty Organizer:
Tami Carpenter, DIMACS
Presented under the auspices of The Homeland Security Center for Command, Control, and Interoperability Center for Advanced Data Analysis (CCICADA).

Abstracts:


Panelist Bios

Clifford Behrens, Telcordia Technologies

Bio:

Clifford Behrens is a Senior Scientist and Director of Information Analysis in Applied Research at Telcordia Technologies. After receiving his doctorate in mathematical anthropology from UCLA, he held a faculty position there, was a Visiting Scientist at IBM and a Visiting Investigator at NASA. He has managed research projects funded by NSF, NASA, DARPA, US Dept. of Education and In-Q-Tel. His research interests include computer and statistical modeling of social-cultural systems, integration of qualitative and quantitative data analysis, knowledge acquisition and ethnographic methods, models of expert opinion and group decision-making, information retrieval and classification, and geospatial analysis.


Patrice Perny, University Pierre and Marie Curie in Paris

Bio:

Patrice Perny is a professor at University Pierre and Marie Curie in Paris, where he heads the Department of Decision, Intelligent Systems and Operational Research. His research is in Algorithmic Decision Theory and involves preference modeling, multicriteria optimization, and computational social choice, as well as decision making under uncertainty and risk. On the one hand, his work addresses elaboration and analysis of new models to describe or simulate complex decision making behaviors; on the other hand, it seeks algorithms that enable fast identification of preferred solutions in combinatorial decision problems. This works comes into play in both decision support systems (for human decision makers) and automatic decision making (by autonomous decision agents).


Rebecca Wright, Rutgers University

Bio:

Rebecca Wright is a professor in the Computer Science Department and the Deputy Director of DIMACS at Rutgers. Her research spans the area of information security, including cryptography, privacy, foundations of computer security, and fault-tolerant distributed computing, as well as foundations of networking. Prior to joining Rutgers, she was a professor in the Computer Science Department at Stevens Institute of Technology and a researcher in the Secure Systems Research Department at AT&T Labs and AT&T Bell Labs. She received a Ph.D. in Computer Science from Yale University, a B.A. from Columbia University, and an honorary M.E. from Stevens Institute of Technology.


Speakers Bios and Abstracts

Tami Carpenter, DIMACS, Rutgers

Bio:

Tami Carpenter is a research professor at DIMACS. Her interests lie in the general areas of planning and optimization, and she is of the belief that her life needs more of both. Prior to joining Rutgers she was Senior Scientist and Director of Network Models and Algorithms Research at Telcordia Technologies. Her duties include acting as faculty coordinator for the CCICADA fellowship program, which enables her to live vicariously through the successes of her array of fellows. She received her Ph.D. in Operations Research from Princeton University, M.S. in Mathematics from Carnegie Mellon, and B.A. in Chemistry and Mathematics from University of North Carolina.


Nina Fefferman, DIMACS & Ecology, Evolution and Natural Resources, Rutgers

Title: Exploring the Role of Behavior in Infectious Disease Dynamics: Mathematical Insights from World of Warcraft and other Virtual Worlds

Infectious disease passes from person to person, from friend to friend, from parent to child, from shopkeeper to customer. Basic social interactions, necessary in every day life, can suddenly become themselves life-threatening in outbreaks of deadly disease. One of the fundamental tools in understanding how diseases will spread, and how that spread will affect society, is mathematical modeling! These models rely on assumed understanding of how people will (possibly) change their behaviors in the face of an outbreak - it turns out that these assumptions can compromise the accuracy of the model predictions.

In 2005, an accidental plague unleashed in the game world, "World of Warcraft (R)" (by Blizzard Entertainment, Inc.), provided a first glimpse of how scientists might be able to exploit these virtual game worlds to study how people react socially to communal threat from infectious disease. We will discuss what current mathematical models of disease spread can predict about disease, and how these virtual games may be able to help us all plan for global pandemics.

Bio:

Nina Fefferman is an assistant professor at Rutgers University in the Department of Ecology, Evolution and Natural Resources and an assistant research professor at DIMACS. Her research lives at the interface of the biological and mathematical sciences and usually falls into one or all of three categories: epidemiology; evolutionary and behavioral ecology; and conservation biology. Sometimes these interests collide, leading her to study the effects of animal behavior, ecology and infectious disease epidemiology on one another. She models disease in both human and animal populations, and is interested in how disease and disease-related behavioral ecology can affect the short-term survival and long-term evolutionary success of a population. Mathematically, this often involves the study of "complex systems" in which each component is relatively simple but when taken together they can interact to create highly organized systems and incredibly complex behaviors. She received a Ph.D. in Biology from Tufts University and holds Master's and Bachelor's degrees in Mathematics from Rutgers and Princeton respectively.


Brad Greening, Ecology, Evolution and Natural Resources, Rutgers University

Title: Optimal Resource Allocation and Evacuation in Urban Crises

We will discuss a model that is currently being worked on in collaboration with researchers at the Centers for Disease Control and Prevention to model public health responses to crisis events in urban centers, specifically in the case of an extreme heat event or an epidemic of infectious disease. In such cases, it is possible for a city to have a spike in individuals seeking medical treatment such that the existing facilities are inadequate to handle all cases. One response of public health officials would be to convert public buildings such as schools, libraries, or sports stadia for use as temporary medical facilities to supply treatment to the overflow. We employ individual-based modeling to examine which facilities would be the best options for such conversions, as well as how to route the population to these facilities in such a way as to minimize loss of life as well as maximize ease of transmission and public understanding of the routing directions.

Bio:

Brad Greening is a Ph.D. student in the Ecology, Evolution, and Natural Resources Department at Rutgers University. He received a B.S. in Computer Science with minors in Mathematics and Music from Rutgers University - Camden. His research interests focus on applying his background in theoretical computer science to answering applied questions in the field of evolutionary sociobiology. He has spent a summer interning at the Centers for Disease Control and Prevention in Atlanta working primarily on modeling efforts for emergency preparedness. Brad is a recipient of a DHS Career Development Fellowship.


Cindy Hui, Industrial and Systems Engineering, RPI

Title: Modeling the Diffusion of Actionable Information in Social Networks

The focus of this research is on studying the diffusion of actionable information in social networks, where the network structure may change over time as the result of the information flow. In addition, the information being spread requires a decision to be made and an action to be taken. An agent-based model for studying such processes is presented. The framework incorporates the concept of trust between individuals and information sources and captures decision-making processes and behaviors at the individual level. We apply the framework to simulate the spread of evacuation warnings in a large-scale wildfire scenario. Through simulation experiments, we explore the effect of network structure in terms of distribution of trust and tie strength in an evacuation setting.

Bio:

Cindy Hui is a Ph.D. candidate in the Industrial and Systems Engineering Department at Rensselaer Polytechnic Institute. She recently successfully defended her dissertation and will be receiving her doctoral degree in Decision Sciences and Engineering Systems in May 2011. She holds a dual B.S. in Mathematics and Computer Science and a M. Eng. in Operations Research and Statistics, also from RPI. Her research interests include model development and simulation, data analytics, network analysis, and social networks.


Scott Kulp, Computer Science, Rutgers

Title: Visualizing Blood Flow through the Heart

After a heart attack, the movement of the heart walls during the cardiac cycle may change, affecting the motion of blood through the heart. This could potentially lead to an increased risk of clotting, and as a result, stroke. In order for doctors to determine a patient's risk, we are working on methods to generate an animated 3D model of a beating heart from CT data, use this animation to perform a physical simulation of the fluid flow, and finally visualize the simulation results.

Bio:

Scott Kulp is a Ph.D. candidate in the Computer Science Department at Rutgers University. He received his B..S in Computer Science and Mathematics in 2008 from Ursinus College and his MS in Computer Science from Rutgers in 2010. He has also spent several summers as an intern at the Department of Defense, working on such projects as topic detection of conversational audio speech and simulations of iris tissue deformation to improve iris recognition performance. Scott is a recipient of a DHS Career Development Fellowship.


Mor Naamon, School of Communication and Information, Rutgers

Title: Network Science and Social Science on Twitter

Online social networks, articulated through many of the today's popular Web applications such as Facebook and Twitter, provide people with ways to create connections, seek support, share ideas, form and maintain relationships, and receive information. In these environments, the structure of a user's "network neighborhood" (ego-centric social network) is likely to be correlated with various aspects of the user' activity and presence in the service. In particular, we present work that uses Twitter to examine the connection between the structure of an individual's social network and a) their emotive communication patterns, and b) persistence of ties in their network.

First, we investigate the connection between individuals' social sharing of emotion and the mathematical properties of their networks on Twitter. We look at emotions that are prevalently expressed by an individual, and correlate those with structural ego-centric network properties. Our analysis suggests that expression of emotion co-varies with users' network properties, and the expression of emotion in directed interactions between users plays an important role.

Second, we investigate how network properties impact tie persistence on Twitter. Building on social theories such as strength of ties, embeddedness, and status, we examine how network structure influences the breaking of ties between individuals in Twitter's directed social network. We investigate this "unfollowing" phenomenon using a large set of Twitter edges, and the persistence or disappearance of these edges after nine months. Our analysis suggests that structural properties of the network have a significant effect on the persistence of ties and unfollowing activity on Twitter. Mor Naamon, Professor, School of Communication and Information, Rutgers

Bio:

Mor Naamon is an assistant professor at Rutgers University School of Communication and Information, where he runs the Social Media Information Lab. His research interests include social media and multimedia information systems, examining the technical, social and human aspects of information. Prior to joining Rutgers, Mor worked as a research scientist at Yahoo! Research Berkeley, where he led a team of research engineers and interns investigating the future of mobile and social media technology. Mor received a Ph.D. in Computer Science from Stanford University. His research in the Stanford Infolab also focused on digital media, and in particular the management of digital photographs. Mor is a recipient of a NSF Early Faculty CAREER Award, a Google Research Award, a Nokia University Collaboration Award, and three best paper awards. In previous careers, Mor was a professional basketball player as well as a software developer and a college radio DJ. In subsequent careers, Mor hopes to be a professional backpacker and traveler. Find out more about Mor at http://mornaaman.com.


Brian Thompson, Computer Science, Rutgers

Title: Detecting Anomalous Activity in Computer and Phone Networks

Detecting Anomalous Activity in Computer and Phone Networks In the last decade we have seen unprecedented growth in the capability to collect massive amounts of data, especially in the cyber domain. Gigabytes of data from communication networks such as cell phone, email, and Internet traffic are captured every second. Our goal is to use mathematical and computer science tools to automatically detect anomalous activity, such as computer viruses or email phishing scams. A communication network can be modeled as a graph, where nodes represent people or computers, and an edge signifies a relationship such as "friends". However, communication networks have a highly dynamic nature - that Alice and Bob are friends says nothing about the frequency or regularity of their communication. To analyze communication in a network, we first build a model based on temporal patterns across each edge, which we call the REWARDS (REneWal theory Approach for Real-time Data Streams) model. With a little help from statistics, we then define a quantitative measure of "anomaly" in an arbitrary subgraph. Finally, we develop graph algorithms to efficiently identify subgraphs with the most anomalous behavior.

Bio:

Brian Thompson is a Ph.D. candidate in the Computer Science Department at Rutgers University, under the guidance of Professors Danfeng Yao and Muthu Muthukrishnan. He received a B.S. in Computer Science with a double major in Discrete Math and Logic from Carnegie Mellon University. His research interests are rooted in algorithms, combinatorics, graph theory, and cryptography, and his projects have included anomaly detection in communication networks, anonymity in data publishing, and privacy and security in outsourced databases. Brian is a recipient of a DHS Career Development Fellowship.


Previous: Program
Workshop Index
DIMACS Homepage
Contacting the Center
Document last modified on April 8, 2011.