At the Service of Society: Rutgers School of Arts and Sciences


Rutgers University is one of America’s oldest
universities. It was founded in 1766, one of the original eight colonial colleges.
The DIMACS center is a little bit younger than that. It was founded in the late
80s through a National Science Foundation nationwide competition for science and
technology centers. CCICADA Center is even younger than that. It was founded in a
nationwide competition for the Department of Homeland Security
University centers of excellence dealing with data science. This is an academic
institution we tried to develop general methods that are usable to solve
problems other than the ones we start with. So in our case we’ve worked with
evacuation modeling and inspection modeling at stadiums and now those have
been applied more recently too large transportation facilities like the bus
terminal in New York City and we’ve used the same ideas to study inspection of
containers as they arrive at the ports. So here we go from developing ways to
solve one problem to using math and computer science and statistics to solve other
problems. My research broadly focuses on developing computation models of human behavior and how humans act and interact in high-density situations such as
crowds for example. One key facet of this research is to ensure that these computation models are grounded in reality
and are able to emulate real human behavior. To facilitate this process we
base our computational models of real empirical observations of human crowds. We
are collaborating with a large number of faculty at Rutgers University to
develop an end to end solution for reconstructing simulating and analyzing
and optimizing crowd dynamics at the Port Authority Bus Terminal. The first
step in this process is to use state of the art light-up technology to reconstruct
the entire Port Authority Bus Terminal. We’re then using advanced computer vision
technology to be able to reconstruct the crowd and then finally using this
reconstructed data as ground truth we are developing computational models to
be able to simulate the crowd in these virtual models. So we can use these tools to
ask what if questions. For example how would the crowd will be impacted if we were to change
the layout of the building of the Port Authority Bus Terminal and this allows
us to even facilitate the design and architectural layout of the bus
terminal without running expensive and often prohibitive human subject
experiments even before the building was actually constructed. So at CCICADA we’ve worked on developing
best practices for stadium security. This has been for different DHS agencies
especially the Office of Safety Act Implementation who deals with liability
protection for a large entertainment venues that have effective
counter-terrorism measures. One of our projects regarding the stadium security
has been in developing mathematical model and simulation for venues to study
their security screening process. At one particular stadium venue they were
considering changing their security screening procedure from metal detecting wands to walk through metal detectors and so we were able to actually use our simulator tool that we
developed at CCICADA to let them see how it would impact their patrons through-put rate and
it turned out that actually their initial idea of how many metal detectors
to acquire were actually far too few so, in the end they had to defer their plan and acquire much more
metal detectors down the road. One of the things that we’ve been researching over
the last year or so is walk through metal detectors in particular. So we’ve done a
lot of experiments with those and this summer we’ve had two undergraduate
students come in through our REU program, Research Experience for Undergrads, and
they’ve been helping me work with these experiments and so the experiments look
at how these metal detectors work indoors and outdoors, on what are the settings you know how did they perform with
people walking through them with different mental test objects. So you
look at different heights, different orientations on someone’s body and how
does that impact detection depending on how you even walk and we are presenting
those results to different sports stadiums and even here at Rutgers to the Rutgers police. It’s an exciting project. DIMACS is the Center for
Discrete Math and Theoretical Computer Science. We were founded in 1989 as an
NSF science and technology center. We continue with our mission to catalyze
and conduct research and education in discrete mathematics, theoretical
computer science, and their related areas: algorithmics, modeling, statistical analysis and so on. My own personal area of research centers a lot around privacy and security and a recent
project that we were involved in was around modeling of human mobility so the
idea is that it’s very useful to know where people concentrate, how they moved
throughout the course of the day, not just as individuals but as a population.
Starting off in their homes in the suburbs, moving into the city to work of course
is a very typical pattern but there’s all kinds of other patterns as well and
our work on human mobility builds on some work with our partners at AT&T Labs
and Princeton and elsewhere looking at how you can take data from cell phone records where
you’ve got just locations, this phone was at this cell tower, made a phone call or made a text, and so you have a very rich data source that can be used to
understand the patterns of how people follow. Our work in particular and our
contribution to that work was to add a layer of privacy in there of
course, of coursse cell phone data has a lot of privacy concerns as well as concerns about
proprietary-ness of the data so a cellphone provider is typically not going to want to
release their data both because they have promises to their customers but
also because there are business secrets in that information and so our work
actually in a very mathematically principled way and sort of what’s called
differential privacy as a layer in the modeling procedure where we reduce the
impact any one individual’s data has on the final outcome and then we did
some experimental work to validate that we were still getting good and useful
results. Rutgers is really at the forefront of work in homeland security and we are
recognized in all the work we’ve already done. And it’s very clear that there’s a lot more to do. But we start from a great place.

Leave a Reply

Your email address will not be published. Required fields are marked *