EDUCATION Projects Year 1
CCICADA Project 1 - Data Sciences Summer Institute
Project Leader:
Nancy Komlanc, PI, University of Illinois at
Urbana-Champaign
Participants:
Dan Roth, University of Illinois at
Urbana-Champaign
Asamoah Nkwanta, Morgan State University
Abdul-Aziz Yakubu, Howard University
Project Description and Goals: DSSI Flyer
The Data Sciences Summer Institute (DSSI) is a keystone of our
education program. We believe that nothing comparable exists in the
nation. DSSI was initially developed at the MIAS Center at UIUC to
encourage computer science (CS) students in universities with small
research programs, [articularly those that are minoroty serving, to
pursue graduate studies and to expose them to national labs. Its
goals are: (1) to provide specific and substansive training for a
new generation of experts in the field; (2) to promote graduate
study in the broad field of information science; and (3) to provide
and intellectual community where participants at all levels of
expertise come together in an environment of collaboration. This
comprehensive education experience is designed to increase
participation in the study and practice of topics of interest to
CCICADA. It will be based at UIUC, with participation by faculty
from all of the CCICADA institutions and EPSCoR states. We work
closely with our oartner MSIs who are given funds to send students
to the DSSI.
It is our intention that the curriculum developed in the DSSI
will be a first stop for anyone whose career - whether practical or
research oriented - is based in knowledge discovery and visual
analytics. DSSI's 6-week residential summer programconsists of a
4-week training session, during which undergraduates, graduate and
participating faculty are in residence at UIUC, coinciding with a
4-week research session, the first half of which coincides with a
training session.. For broader and more rapid impact, we will
invite psrtner institutions and national labs to rely on these
course materials as they establish research or training groups of
their own. Our program specifically aims to attract, engage, and
nurture students and faculty whose talents or research interests
fall in one of the following broad technical fields of information
Science: 1) Databases and Information Integration; 2) Natural
Language Processing and Information Science; 3) Computer Vision,
Image Processing and Information Visualization; 4) Information
Retrieval and Web Information Access; and 5) Machine Learning and
Data Mining.
DSSI Component Activities:
The DSSI summer program is integrated with the center's research
goals by weaving together the following four coordinated
activities:
Mathematical Foundations of Information Science:
This course consists of a reigorous treatment of Probability, Linear
Algebra, Data Structures and Algorithms, Optimazation, Data Analysis,
and Clustering and Visualization. applications and illustrative
examples are drawn from the broad technical fields in information
science.
Application Courses
To immerse students in the practice of knowledge discovery
and prepare them for research project in their area of interest, we
integrate a sequence of technical tutorials. Researchers from the
CCICADA centersteach courses, making the content up-to-the-moment
andkeenly relevant. Each undergraduate enrolls in at least three of
the five tutorials offered. Tutorials in DSSI 2007 and 2008 at our
predecessor COE MIAS were: Databases and Information Integration;
Natural Language Processing and Information Extraction; Computer
Vision, Image Processing and Visualization; Information Retrieval and
Web Information Access; Machine Learning and Data analysis. New
tutorials relating to data visualization and large-scale computing
will be offered in the 2010 program and beyond. Material presented in
these tutorialsis electronically available for the broad online
distribution. Attendance by DHS, the national labs, and CCICADA
partners is welcomed.
Summer Research Program:
To complete the educational experience for students and faculty,
we offer a summer research program whose primary goals are to give
students a taste of the exhilaration that comes from academic
discovery, and to give professional researchers an opportunity to
pursue their interests with a well-trained staff. Undergraduates who
have taken the foundationscourse, graduate students, and faculty from
CCICADA institutions and researchers from national labs collaborate on
well-defined and appropirately scoped research problems.
Expert Speakers Series:
We augment the students' training through the speaker series which
features invited distinguished researchers and practitioners who
report on their experiences via public lectures that are simulcast to
partner institutions and national labs and archived for future
access.
Accomplishments to Date:
The Data Sciences Summer Institute (DSSI) has been taken to the
next level using an online registration system which includes an
application review system, a successful national promotion campaign
with the help of our partner CCICADA institution rutgers, and
marketing materils sent to over 700 university career centers, CS
department heads, Data Science resesrchers, CS professional
association list serves, CS students as well as posting on numerous
websites, blogs, press releases, and yahoo groups. These efforts have
provided a very large, well-qualified pool of over 160 students
applying from MSIs and other U.S. universities.
Work to be done in Rest of Year 1:
The DSSI is held May 24 through July 2, 2010. Before the end
of April all the Expert Guest Speakers are confirmed, the DSSI student
roster confirmed and confirmation notices sent, all UIUC on-site
housing contract signed, all computer labs readied, program
evaluations created, and research topics finalized.
Interfaces wit Research Tasks/Projects at CCICADA:
We are calling upon CCICADA researchers from a variety of CCICADA
institutions to give guest lectures and provide mentors and will plan
future themes of DSSI around major CCICADA research themes.
Connections to work at other COEs and VACCINE:
Researchers from other COEs and VACCINE will be guest lecturers and
faculty mentors as DSSI moves forward. The Education Director for
DSSI, Nancy Komlanc, was on the education panel at the annual DHS
SUMMIT in D.C. in March 2010, discussing building relationships with
MSIs. That experience openned up opportunities to network with other
COEs' Education Representatives for the comming year. The data science topics covered in DSSI are fundamental tools in
all areas of homeland security. Proposed Wwork in Year 2: A new DSSI initiative will offer Data Science workshops to
corporations within a 50 mile radius of the University of Illinois
campus. The short list of corporations we would target are:
Caterpiller, John Deere, State Farm Insurance, Country financial
Insurance, GE, etc. Currently, the UIUC AI Group hosts several
seminars per semester through the Artificial Intelligence and
Information Systems Seminar (AIIS) seminar series. This seminar
series brings in AI and Data Science experts from around the country
for discussions on current and future topics giving attendees
opportunity to schedule private meetings with the speaker. The UIUC
computer Science Department already has Corporate Relations personnel
in place to work with us to make these corporate contacts.
CCICADA Project 2 Reconnect Program
Project Leader: Project Faculty & Participants: Student Assistants: Project Description and Goals:
The CCICADA/VACCINE Summer Reconnect Conferences expose faculty
teaching undergraduates to the role of the mathematical and computer
sciences in homeland security by introducing them to current research
topics that are relevant for classromm presentation. They also offer
an opportunity to researchers in government or industry to learn about
recent techniques. Toics are presented in a week long series of
lectures and activities led by leading experts in the field.
Participants are involved in both research activities and in writing
materials useful in the calssroom or to share with their colleagues,
with the possibility of ultimately preparing these amterials for
publication. in our Educational Modules Series. These conferences
offer the opportunity for junior faculty as well as mid-level and
senior faculty and government and industry professionals to advance
and research questions in a new area of the mathematical sciences.
Participants acquire materials and gain ideas for seminar
presentations and for undergraduate research projects and have the
opportunity to network with people from a variety of backgrounds. For
our first Reconnect, the audience was primarily college faculty. For
our second, we are getting a number of homeland security
professsionals. Accomplishments to Date: Reconnect 2009 This Reconnect Conference was truly a joint CCICADA-VACCINE
effort. It introduced participants to basic concepts in visual
analytics, an emerging field that integrates visualization with data
analysis to facilitate analytical reasoning. Visual analytics
envisions a discovery process in which humans, algorithms and
visualizations interact to solve problems. Program participants gain
an understanding of how the various componenents of visual analytics
(human, algorithmic, and visualization) can work together to solve
large and complexreal-world problem in such areas as homeland security
and bioinformatics. Each participant used a variety of algorithms and
visualizations; each worked with realistic data sets; and assigned
readings as well as hands on use of several commercial and open source
tools. Lectures were given by Georges Grinstein of UMass Lowell and
John Stasko of Georgia Tech. One interesting outcome of Reconnect 2009 is that Professors Sam
and Ahlam Tannouri from our partner Morgan State University, Who were
"students" at the conference, got very excited about visual
analytics. Their presentation at the CCICADA center-wide retreat in
March 2010 reflected this excitment and their follow -up work has led
to their becoming engaged in our research projects 8.5 and 10.5, as
noted earlier. Conference Participants: Reconnect 2010 Automatic identification and extraction of desired information from
natural language text is increasingly used as a way to improve
general-purpose search and has a range of applications, and for the
intelligence community. The input is one or more texts in the domain
in question, and the output is a database containing just the desired
fields of information, extracted from the source material and
formatted appropriately. Information Extraction (IE) techniques have
been developed sinc ethe early 1980s, and include finite state
technology, pattern-based extraction, and appropriate machine learning
methods. This week-long Reconnect 2010 conference will take
participants from the early, simpler, methods through the modern ones,
and will include theoretical and practical topicsas well as hands-on
exercises using software packeges. The material is very relevant to
the undergraduate classroom and to many applications. The lecturers
are renouned experts in the various aspects of IE and its
visualization, and we have a long history of giving informative,
engaging, and fun lectures. Work to be Done in Year 1: Interfaces with Research Tasks/Projects at CCICADA: The 2009 summer Reconnect workshop topic was Visual analytics and
was organized jointly by Georges Grinstein of CCICADA and John Stasko
of VACCINE, linked to both centers' research work. The 2010 summer
Reconnect workshop topic was chosen to link to research projects of
Dan Roth and Ed Hovy at CCICADA and John Stasko at Georgia Tech. Connections to work at other COEs and VACCINE: Proposed Work in Year 2: We plan to hold reconnect regularly once or twice a year,
collaborating with VACCINE. Plans for a repeat of the 2009 Visual
Analytics Reconnect, this time to be held at Georgia Tech, are being
discussed with VACCINE and have been delayed due to the availability
of John Stasko, one of the principal lecturers. We will selct a topic
for a new data analytics Reconnect to be held in 2011, develop plans
for it, and aim to hold it in June 2011. Publications: A number of participants drafted "modules" based on Reconnect
2009and others drafted research papers. These modules and research
papers are being updated. The "modules" aim at bringing homeland
security research or methods relevant to homeland security research
into the undergradute classroom. We aim to elaborate on the best of
these and work them into the CCICADA modules program (see Education
Project 4)
CCICADA Project 3 on on Research Experiences for
Undergraduates
Project Leader:
Gene Fiorini, Rutgers University Project Participants: Tamra Carpenter, Rutgers University Project Description: The CCICADA Research Experiences for Undergraduates (REU) program
offer one-on-one research project experience to undergraduates under
the guidance of CCICADA researchers. The program leverages a large
REU program that has been run out of the DIMACS Center at Rutgers
since the early 1990s. CCICADA aims to support approximately fice
students from aroundthe country each year and in addition, teams with
our MSI partner institutions to include their students in REU. In
summer 2009, we hosted students under our previous DHS grant for the
Center for Dynamic Data Analysis (DyDAN), with partial support from
CCICAD. In summer 2010, the program will come fully under CCICADA
support. The program begins with eight weeks of intensive work during the
summer, students are strongly encouraged to continue their projects
during the following academic year. Participants will have a faculty
member as a mentor, in most cases working one on one with their
mentor. Housing will be provided as well as a modest stipend and
travel support. Students are required not to commit to other
activities (e.g., courses) during the REU program. Applicants to REU should be undergarduates with a minor in Computer
Science, Mathematics, or a closely related field. They should be
current juniors, although sophmores with exceptionally strong
backgrounds will be considered. In exceptional circumstance, seniors
will be allowed into the program if they will not have graduated by the
time the program begins. Preference will be given to students who will
continue their research projects during the academic year, under the
dirction of either their mentor or a faculty member from their home
institution. Mentors are all volunteers and are not compensated for their work. Accomplishments to Date: The summer of 2009 REu program was a great success. Some of the
student projects are summarized here. The projects listed included
some mentored by CCICADA faculty and conncting to CCICADA research,
but funded by other sources, including our predecessor COE DYDAn. In
each case, the undergraduate student is indicated with an
asterisk*. REU Student Curtis McGinity* (Tulane University), Minge
Xie, Rutgers University Department of Statistics and
E. A. Elsayed Rutgers University Department of Electrical
Engineering Optimum Strategies of Container Inspection at Port-of-Entry Ever-increasing globalization has significantly increased the
number of cargo containers being transported internationally.
additionally, it is estimated that shipping containers account for
approximately 95% of the world's international cargo in terms of
value. While most contain food, equipment, raw materials, or other
commodities, some percentage of containers carry undesirable items
such as drugs, weapons, chemicals, and possibly even nuclear
materials. As such, it is imperative to develop strategies for
detecting contraband that can act as a viable deterrent against
illegal shipping while leaving the fluidity of the dhipping industry
uncompromised; it would simply be infeasible to completely inspect
even a statistically significant sample of the population of
containers. Previous work at Los alamos and by CCICADA reserchers models the
container inspection process as a set of n independent sensors,
each with a corresponding threshold value T, that classify each
container as "good" or "bad." Then, givena specified Boolean decision
function, and overall decision of "clear" or "suspicious" (i.e.,
requiring manual inspection) is made for each inspected container.
This work detailed in a multi-objective optimization approach to
determine the optimal sensor threshold levels and sensor sequence
while incorporating inspection time, misclassification and measurement
error. Together with his mentor E. A. Elsayed and one of the
CCICADA's PhD students, Tsvetan Asamov, Cutis McGinity (Mathematics,
Tulane University) employed a randomized flow network through various
sensors so that a given container would undergo a completely different
inspection process upon re-entering the sensor network. In this way,
the intrinsic unpredictability of the inspection process acts as a
deterrent to would-be attackers. the team realized that, for a given
set of thresholds Ti, the problem of finding the randomized
inspection policy could be formulated as a generalized linear
programming network flow problem. REU Student Alexander Crowell* (Rutgers University),
Danfeng Yao, Rutgers University Computer Science Department Detecting Drive-by-Downloads Using Human Behavior
Patterns A previous REU project mentored by Dr. Yao investigated the
similarities and differences in HTTP periodicity between botnet
command and control traffic and legitimate web server traffic in order
todetect running botnets. In this project, Alexander Crowell and Dr
Yao sought to apply a similar approach to the problem og
drive-by-downloads, where malicious web page installs software on the
user's computer without their permission, to investigate the
similarities and differences between user-permitted downloads and and
malicious ones and investigating how difficult it is for the malicious
sites to fool our detection mechanisms, alexander and Dr. Yao sought
to create a reliabe system for detecting and preventing drive-by-downloads.
REU Student Chinua Umoja* (Morehouse College),
William Pottenger, Rutgers University DIMACS and Computer Science Department Higher Order Learning Traditional machine learning approaches make the assumption that
assume instances are independent and identically distributed (IID).
Models constructed under the IID assumption are termed first-order
because in general they only leverage relationships between attributes
within instances (e.g., co-occurrence relationships). Thus
classification of a single instance (of previously unseen data) is
possible because no additional context is needed to infer class
membership. Such a context-free approach, however, does not exploit
valuable information about relationships between instances in the
dataset. This project built on the novel framework for learning that, unlike
approaches that assume instnaces are IID, leverages implicit
co-occurrence relationships between attributes and instances. These
implicit co-occurrence relationships are termed higher-order
paths. Attributes (e.g., words in documents in text collections) are
richly connected by such higher-order paths, and the model built by
the faculty mentor's higher order learners exploit this rich
connectivity pattern. The project built on work to-date on both
supervised and unsupervised learning approaches including Higher Order
Naive Bayes, Higher Order SVM, Higher Order Classification Based ARM
and Distributed Higher Order ARM. It also contributed to a framework
that leverages human-computer interaction entitled Distributed
Interactive Higher Order Privacy Enhancing Knowledge Discovery (DI
HOPE KD). REU Student Shyretha Brown* (Jackson State University),
Nina Fefferman, Rutgers University Ecology and DIMACS Optimal Locations for Evacuation Facilities during Heat
Events Extreme heat events overtax energy and water needs of cities,
eventually compromising infrastructure and safety of homes, offices,
and public facilities. Increased incidence of heat stroke,
dehydration, cardiac stress and respiratory distress are commonly
resulting health problems. These can be especially serious among
elderly or juvenile populations. Under severe enough conditions,
evacuation to controlled environments can be the best means of
ensuring the ontinued well-being of the population. However, the
determination of optimal placement of evacuation facilities can be
difficult. Facilities must be able to amintain energy and water
supplies and suffiient, hygenically-maintained space for displaced
persons. They must be able to manage incoming supplies of food and
potable water despite the heat-related increase in the dangers of food
spoilage. Further, the populations at greates health risk from heat
events are aso those least able to travel long distances, requiring
consideration of spatial demography for the area being served by the
facility. Easy access to healthcarewill also be of great importance,
whether that should ultimately include planning for onsite care, or
ensuring nearby access to hospitals capable of handling the increased
patient load. Careful planning for the locations chosenfor evacuation
faciities may be of critical importance to ensuring minmal health
impact during heat events. This project researched some challenges in
this area which cut across disciplines and involve spatial demographic
distribution of vulnerable populations, probabilistic mixed integer
programing methods, and other aspects of "location theory." REU Student Andrew McConvey* (University of Notre Dame),
DHS Summer Research Team Student Joshua Smith* (Morehouse
College), Nina Fefferman, Rutgers University Ecology and DIMACS Sensitivity of Entropy Measures in Biosurveillance Early work by the mentor and her collaborators and students has
demonstrated that the use of information theoretic entropy can detect
outbreaks in streaming disease incidence data earlier than other
methods. However, this work has required human-based decisions using
prior knowledge about disease specific processes. this project
explored general preprocessing parameters for disease-blind
sensitivity. REU Student Emmanual Williams* (University of
Maryland-Baltimore County),Nina Fefferman, Rutgers University
Ecology and DIMACS, Tamra Carpenter,Rutgers University DIMACS Allocation of Monetary resources in HIV Infected
Communities In many improverished nations, monetary resources to treat HIV are
severly limited, so communities are unable to treat all of the
infected population. While much research focuses on specific treatment
strategies or more macro-scale modeling, recent research is also
looking at finer modelsof equitability and economics. This prject
developed and refined a mathematical epidemiological model that ties
community economics with HIV progression. Through computer
simulation, the model was used to gain intuition and potentially
inform policy decisions on how to allocated treatment resources. this
model was then refined to capture features of populations in specific
areas (e.g., sub-saharan Africa) and used to explore the impact of
different treatment strategies, with the goal of identifying those
that best assure the stabiity and long-term survival of the
community. REU Student Jeffery Truman* (University of Arizona),
Paul Kantor, Rutgers University School of Communication,
Information and Library Studies Game Theoretic Aspects of Homeland Security The effort to protect our country from the threat of smuggled
wepons is a non-zero sum game, in which the opponent's estimates of
the value of a success are different from our own estimates of the
cost of failure. The problem can be formulated as an "Inspector
Game." This project examined this problem in the context of an
ongoing research program on the problem of optimal detection of
nuclear threats a t borders, and within the country, REU Student Elise Aspray* (Rutgers University), REU
Student, Diane Render*, (Albany State University) and
Midge Cozzens, Rutgers University DIMACS A game Theoretic Approach to Finding Effective Anti-Terrorism
Strategies Consider a game in which the players are coalitions formed by one
or more government groups (countries. Any number n of
coalitions can playat one time and the coalitions can vary in size.
If the game contains four items of value/utility: Lives , money, chits
in heaven, and reputational capital, how will the game play out if
only two procedures, actions (be agressive, do something) or no action
(do nothing), are allowed by each coalition? This problem considered
Nash Equilibrium (complete information), Bayesian Equilibrium
(incomplete information taking into account the probabilities of
coalitions choosing each action) senarios for non-sequential games
(zero-sum games and non-zero sum games) as well as sequential games
(coalition A takes and action, then coalition B reacts, then A, then
B, etc.). Some issues considered included: How to define lives - lost
or saved? How to quantify chits in heaven? Is the quantitative
definition of reputational capital for coalitions the same? DHS Scholar John Kim (Massachussets Institute of Technology),
Nina Fefferman, Rutgers University Ecology and DIMACS Self-organizing Social Networks Populationsin which individuals chose their friendships have been
shown to be able to converge to stable networks. This project
explored whether or not certain affikiation preferences can be proved
to converge vs. diverge under different network measures of success.
the project built on previous work by Fefferman and her colleagues
examing the impact of disease dynamics and response behaviors on the
organizational success and stability of a self-organizing population
of individuals. John's research provided a theoretical analysis of
boundary conditions for parameters of switching rates in dynamic
networks that couls yield previously empirically observed outcomes in
the transmission of infection across social networks.
Work to be done in Rest of Year 1: Plans are well underway for the 2010 REU program. We receive over
300 applications for the joint DIMACS-CCICADA REU program. We have
selected five of these students to be funded by CCICADA to work with
CCICADA faculty members on projects of direct relevance to our work
program. In addition, Morgan State, Howard University, and Texas
Southern University have each selected a student to participatein the
REU program, also under CCICADA funding, and Tuskegee is in the
process of selecting a student. In addition, we have funding from NSF
to supprt the large number of talented applicants, many of whom will
also be mentored by CCICADA faculty or industry researchers on
projects of direct relevance to homeland security. The projects
described above are representative of the types of projects that will
be offered in 2010. the full list of 2010 projects is provided on our
REU website: http://dimacs.rutgers.edu/REU/. Interfaces with research Tasks/Projects at CCICADA: The REU program is designed to involve undergraduates in CCICADA
research. Connections to work at other COEs and Vaccine: Several of the REU projects are built on joint CCICADA projects
with other COEs Homeland Security Motivation: Connections with Homeland Security Practitioners: The REU program has a weekly seminar and we hope to invite homeland
security practitioners to be presenters at the seminar. Proposed Work in Year 2: Each REU program spans two CCICADA fiscal years, so the summer of
the 2010 program will continue in year 2. the summer 2011 program
will span years 2 and 3 of CCICADA. Unless we receive more funds in a
second installment, the number of DHS-supported REU students for
summer 2011 will be limited to two rather than our preferred target of
five.
CCICADA Project 4 on K-16 Module Development
Project Leader: Project Faculty & Participants: Project Description and Goals:
We are developing four to ten 5-8 day modules for middle school,
high school and /or college classroom, and in some cases for use with
homeland security personnel. Topics, with applications to homeland
security will be chosen in the general fields of data analytics and
visual analytics (Through collaboration with VACCINE) within the mathematical and computer sciences areas. The objective of this project is to bring current research in the area of data analysis and its applications to homeland security to students K-16 and homeland security professionals in an appealing teachable format. Authors will be selected from DHS Center of Excellence Participants or those suggested by the participants.
Accomplishments to Date: We have worked to establsh the topics for the modules, identify
authors, and begin module preparation. Steve Miller at Williams College and Margaret Cozzens at Rutgers
are working on a book in Cryptography, with applications to
Homeland Security for undergraduate Liberal Arts and Criminal
Justice students, who are likely to pursue careers in the FBI, CIA,
Homeland Security Agencies, etc. At least three modules will be
pulled from the book, for use in high schools, and community
colleges. A module on Tomography and Applications to Homeland Security (food
safety and structural soundness) is under way as are two modules from
the Crytography book. Work to be done in Rest of Year 1: Work will continue on the module selection process, the book and
modules from it, and the tomography module. We expect a full outline
of this first module by the end of year 1 as well as a preliminary
outline of the other two modules. Interfaces with Research/Tasks at CCICADA: Many module topics are going to be chosen to link with research
projects. Connections to work with other COEs and VACCINE: We expect that VACCINE will coordinate with us a module developemnt
of module topics, we got a late start in the writing process. We have
a very good start on three modules, two from the cryptography book and
one on tomography. the funds for the module projects are for honoria
for th authors, publication costs, and a workshop that will present
modules to teachers who will use them. The workshop is going to be
held later than originally planned because we started on the modules
later than expected and so they will not be ready for a teacher
workshop in the first year. Thus, the funds for the workshop and some
of the publication costs will need to be arried over. We also saved
on honoria since one of the authors of the two modules as Midge
Cozzens, who is being paid a salary. We estimate that there will be
around $24,000 left to be carried over to year 2. Proposed Work in Year 2: In the first half of year 2, we expect to have developed two
modules from the Cryptography book by Miller and Cozzens, drafted and
pilot-tested the tomography module, and developed an outline for that
module. We will also hold the planned teache workshop at a time that
seems most appropriate to get enough teachers to come. Because most
of the funds for this will be carried over from year 1, we only ask
for $785 fr tghe modules project in year 2. The Reconnect Program (Education Project 2) is also a potential
source of modules. Participants in Reconnect are asked to develop
modules for use in their classrooms. We will identify the most
promising of these and see if we can develop them further for
inclusion in the CCICADA modules series. CCICADA Education Project 5 on Workshops/Tutorials Project Leader: Project Participants:
Project Faculty & Participants: Endre Boros, Rutgers University Project Description: Our goal is to organize at least two workshops per year related to
CCICADA themes, and we will seek to increase this by leveraging other
resources through joint sponsorship. some of the workshops will be
tutorial in nature, aimed at students or homeland security
professionals. Others will be for specialized researchers on the same
topic. Most will be interdisciplinary in nature and will align with
major research areasin data science and its applications. At Rutgers, we are also organizing an onterdisciplinary seminar
series featuring topics of interest to CCICADA researchers and
students from both Rutgers and nearby partner institutions, and
featuring speakers from the homeland secuirty communityy as well as
other COEs and VACCINE. Similar seminar series are being run at some
of our partner institutions. These series allow us to enhance
interactions of various kinds, by inviting speakers from CCICADA
institutions to visit eachother, speakers from other COEs and VACCINE,
andspeakers from the homeland security community. also, the six
CCICADA graduate student fellows at Rutgers are sponsoring a
student-organized seminar series, funded through two Career
Developemnt Grants from DHS. Accomplishments to Date: CCICADA sponsored the following workshops:
Workshop on Mathematical Models for behavioral
Epidemiology: Recent worries about pandemic flu and
bioterrorism threatshave increased the scope of questions to be
addressed by behavioral epidemiology: how would individuals react to
widespread disease exposure risks; how can we expect individuals to
act if offered protective vaccines that themselves carry the risk of
adverse events; and most importantly, how could these individual
behaviors themselves influence the scope of an outbreak? One of the
most interesting and primarily unaddressed questions in this area is
whether or not local behavioral interventions can scale up to affect
nation-wide disease risks. answering these questions for enemic,
epidemic and pandemic disease threats, whether natural or introduced,
will necessitate the creation of new types of investigative models.
to explore such questions, this workshop brought together applied
mathematicians in the area of game theory, graph theory and
operations, mathematical epidemiologists (both human and wildlife),
economists, animal behavioralists, conservation ecologists, medical
sociologists, and public health officials. Our seminar series are off to a great start. By way of example, at
Rutgers, some examples of seminar speakers have Milind Tambe of the
CREATE COE (speaking about game theory for security, lessons learned
from deployed applications); Jeremy Wright from CCICADA partner AT&T
(speaking about detecting changes and anomalies in noisy text
streams); Senelani Dorothy Hove-Musekwa from the African Institute of
Mathematical Sciences (speaking about the impact of stress related
immunosuppression in disease outbreak dynamics); Coast Guard Captain
Todd Gatlin (speaking about Coast Guard operations in the Delaware Bay
and New Jersey coast); Alex Welteof the University of Witswatersrand
in South Africa (speaking about HIV incidence estimation); Janet Kim
of Locheed Martin (speaking about visual analytics research and
development at Lockheed Martin Advanced Technology Laboratories); and
Tina Eliassi-Rad of Lawrence Livermore national Laboratory (speaking
about mixed-membership community discovery). Jack Jarmon organized a global security symposium at the University
of Pennsylvania to address issues relative to counterintelligence and
discuss policy options among national security experts in government,
the private sector, and academia. Work to be doe in Rest of Year 1: The following workshops are planned for the rest of Year 1: Workshop on Statistical Issues in Analyzing Information from
Diverse Sources: today's decision makers in fields ranging from
engineering to medicine to homeland security have available to them
remarkable new technologies for gathering potentially huge amounts of
information from diverse sources. Yet, decision makers are often at a
lossas to how effectively combine and analyze information from
disparate sources. Instead they often look at different sources
separately and invoke relatively as hoc methods for gaining combined
intuition. this approach may lack rigor and fail to harness the full
value of the information at hand. This workshop will explore methods of, theory for, and barriers to
combining data from various sources for improved decision making that
exploits inferences that are typically more efficient and potentially
more accurate than those from any single source. The workshop will
bring together statisticians, applied mathematicians, computer
scientists and policy makers to address issues related to combining
information, and it will disseminate research results in the areas of
model building, Bayesian analysis, incorporation of expert opinion,
meta-analysis, and machine learning, among others. The topics will
encompass statistical and mathematical approaches as well as
computational tools that are related to combining information from
different sources for inference, learning and decision making. Questions of potential interest during the workshop include: - How can we combine information from different studies of
the same type? The workshop will examinetimely and important applications from a
variety of fields. These include medicine and public health, as well
a applications from industry and homeland security. Workshops on Modeling and Mitigation of the Impacts of
Extreme Weather Events to Human Health Risks: Recent unexpecte
extreme weather events are testing our emerging response capabilities,
such as the heat wave in Chicago in 1995, hurricanes Floyd in 1999,
Katrina 2005, and Ike in 2008. the heat wave caused 514 heat related
deaths (12 per 100,000 population) and 3300 excess emergency
admissions. In case of such extreme event, people can move ti
shelters where climate controlled invironment can be provided and
healthcare needs can be met. Without a detailed plan that takes
population demographics and uncrtainty into consideration, there may
not be sufficient accomodation in health centers (such as hospitals,
clinics, and shelters and others) for the population in the target
areas. this workshop will bring together emergency management
disaster management practitioners and the operations research
community to discuss emerging needs and the related current research
issues. focus areas that will be of special interest include:
Statistical analysis of healthcare needs of populations; modeling of
heat events; Facility location and resource allocation; workforce
scheduling. Interfaces with Research Tasks/Projects at CCICADA: Connections to work at other COEs and CCICADA: The proposed year 2 workshop on Adversarial Decision Making
(described below) includes a CREATE researcher (Miland Tambe) on the
organizing committee, and we anticipate that there will be several
additional CREATE participants. Homeland Security Motivation AllCCICADA workshops either deal with explicit homeland security
applications or deal with tools/methods relevant to a variety of
homeland security problems. Connections with Homeland Security Practitioners: The workshop on mathematical modelsfor behavioral epidemiology
featured a speaker from the Cneters for Disease Control (CDC). The
workshop Modeling and Mitigation of the Impacts of Extreme Weather
Events to Human Health includes as a theme exploring the effects of
extreme weather events and we expect it will be of interest to NJ
agencies such as the NJ Department of Health and SEnior Services and
the CDC. Proposed Work in Year 2: The following workshop is scheduled for year 2. Others will be
added as we identify new topics. We will continue to look for new
opportunites to leverage funding by co-sponsoring workshops and
tutorial programs relevant to CCICADA. Workshop on Adversarial Decision Making: Adversarial
decision makingarises in amny situations, including counterterroism,
corporate competition,and federal regulation. Military leaders,
corporate executives and consumer groups regularly make large
investments in the context of intelligent opposition. such choices
typically entail high-consequence outcomes conditional on low
probability events, with solutions drawn from the fields of decision
analysis and game theory. . However, previous work in decision
analysis has largely overlooked adversarial situations - instead, it
has focused on the uncertaintiesassociated with natural disasters and
comcomitant costs. similarly, game theory has largely overlooked
realistic uncertainty sucha s low opponents anticipate and adapt to
eachothers' actions, resulting in a formulation which makes untenable
assumptions about how humans process information and cope with
uncertainty. In particular, it does not account for highly
intelligent opponents, who attempt to maximize their gain based on
game theoretic models of the thinking of the other palyers. These
issues lead to the following three interlocking research questions
that will form the agenda for this workshop: 1) Given a set of beliefs
and partial information about an opponent's resources, intent, and
ability, how can a decision amker choose investments that most
advantage their own objectives? 2) How should one model decision
processes, based on models of other people's decision making, in
practical ways that avoid infinite regress, unrealistic computation,
and simplistic assumptions? 3) Can adversarial risk analysis lead to
usefully different results than those found in classical game theory,
and how may such differences be exploited by decision makers? CCICADA Education Project 6: NAM MATHFest Educational Program
for Moinority Students Project Leader: Project Paticipants: Project Description: The National Association of Mathemeaticians (NAM) is a non-profit
professional organization that aims to promote excellence in the
mathematical sciences, encourages the mathematical development of
underrepresented American minorities, and adresses the
serious-representation of minorities in the mathematical sciences
workforce. NAM achives its goals by focusinng on fice areas:
- Mathematics Education Although the majority og NAM members belong to under-represented
groups of American minorities, a significant number of members
represent a cross-section of the mathematical sciences community.
Membership is oopen to all. NAM's Mathfest is held annually in the
Fall and features a series of speakers, panels, and educational
sessions aimed at undergraduate and graduate students and faculty
nationwide, with an emphasis on MSI institutions. Students give
presentations on their reserch and receive guidance on careers in the
mathematical sciences, applying to graduate schools, and surviving
graduate school. This project is to help design the program for NAM
Mathfest, give it a homeland security emphasis, and assist with
presentations. Accomplishments to Date: We worked with the chair of Mathfest, Professor Leon Woodson, to
design the November 12-14, 2009 meeting held at the University of the
District of Columbia. We rranged for the Director of CCICADA, Fred
Roberts, and the CCICADA DHS Program Manager, Joe Kielman, to give
major presentations. Former REU students at our predecessor DHS
Center, DyDAn, were also lined up to participate as role models who
have gone on to graduate school. Here are the details of the
program: NAM MATHFest Educational Program for Minority Students Organizers: Nathaniel Dean, National Association of Mathematicicans (NAM) Work to be done in Rest of Year 1: We have already begun to plan CCICADA engagement with NAM MATHFest
2010. This work will continue until the end of Year 1.
Interfaces with Research Tasks/Projects at CCICADA: As we continue to work with NAM, we will seek to include CCICADA
speakers from a variety of CCICADA projects. Connection to work at COEs and VACCINE: We had plan to involve VACCINE and NAM MATHFest, but VACCINE did
not have the funds to participate this year. We hope they will in
Year 2. Homeland Security Motivation: Connection with Homeland Security Practitioners: Among the speakers we are hoping to engage in NAM MATHFest in the
future are homeland security practitioners who make use of
mathematical science tools. They would make wonderful role models for
the students attending MATHFest. Proposed Work in Year 2: We will support NAM MATHFest in 2010 in the same way we supported
NAM MATHFest in 2009. However, Funds will be limited unless we
receive a second installment of Year 2 funding. CCICADA Education Project 7: Etoys - and StarLogo - based
Curriculum and Toolbox Development for K-12 Education: Sart-up
Phase Project Leader: Participants: Project Description: We will create a set of ten K-8 computer science activities and
pilot them in collaborating schools. We will also create a hybrid
model(online and face-to-face) course for computer scienc instruction
at the elementary and middle school levels and prepare a research
report analyzing the impact of these modules. We will prepare
professional development materials for pre-service and in-service
teachers. We will apply for external funding for this activity. Accomplishments to Date: We presented Etoys material to children in Washinton, D.C. and
completed online activities "CS4K5" as examples of computer science
activities, including programming, for elementary school children. We
submitted a preliminary response to DHS BAA09-07 in a white paper by
Kathleen Harness, George Reese, Daniel Wendel, Eric Klopfer, Amon
Miller, Karen Brennan, and Michael Resnick. We also submitted a
proposal to "Etoys-CS" to Office of the Vice-Chancellor for public
Engagement at the University of Illinois. Work to be done Rest of Year 1: We received notice on March 23, 2010 that our white paper was one
of those chosen for submission of a full proposal. We will complete
this proposal before the end of May 2010. We are also creating collaborative modules in Etoys and StarLogo.
Translating applications in one program to another gives us a sense of
richness of each tool, and provides the groundwork for our midule
development. Interfaces with Research Tasks/Projects at CCICADA: We will look to research tasks at CCICADA for topics for the
modules. Homeland Security Motivation This project aims at development of the homeland security workforce
of the future. Proposed Work in Year 2: We will be continuing the project in year 2, again without CCICADA
budget. We will pilot an integrated (Etoys/StarLogo/Scratch)
professional development course. If the proposal to the BAA09-07 is
funded, we will be able to do this hybrid online model. In Year 2 we
plan to complete transactions of 4 key elementary and middle school
modules into other programming languages. We plan to pilot the
Etoys/StarLogo module and develop a syllabus for online corse for
an in-service and pre-service teachers. CCICADA Education Project 8: Methodology of Evaluating
Analytics This project coincides with Research Task 8.6 and is described
under that ask.
CCICADA Education Project 9: Development of New Homeland Security
Courses and Homeland Security Curricula Project Leader: Project Description: The development of new homeland security courses and new homeland
security degrees or certificate programs at CCICADA institutions is a
long term goal at CCICADA. It should be noted that this was not
specifically included in the Year 1 workplan and was not given any
specific funding, nor do we propose any specific funding in Year
2. Accomplishments to Date: This semester at Princeton University, Warren Powell is teaching a
course entitled "Optimal Learning." This was first taught in the
spring of 2008 as an experimental one-time-only course; now it is a
permanent course with approximately 60 undergraduates enrolled. Also
in production is a textbook "Optimal Learning" under contract with
Wiley and Sons. As of this writing the book has voer 200 pages, and
Powell anticipates sending it to the publisher in the summer of 2011.
The book is written at an advanced undergraduate level (a clourse in
probability and statistics is required). The development of the
course and the bookbegan with our predecessor center DyDAn and
continues under CCICADA. At Texas Southern University, Lila Ghemri develped a new course
called "Privacy Preserving Technologies." Thge course was developed
as part of a new concentration in information security and assurance,
and aims at providing an understanding and awareness of privacy issues
in the global networked world. it provides an overview of the
different privacy laws in application in the U.S. and globally. It
also presents the use and design of technologies enforcing privacy
aspects. course topics are: Definition of Privacy and Data
Protection, Privacy Principles and Legislation; Privacy Threats and
Risks in the Global Information Society: Profiling Practices,
Elctronic Commerce, Identity Theft, PET for Protecting User
Identities: Protection at Communication Level, Protection at System
Level, Protection at Application Level, PET for protecting Databases:
Inference Controls for Statistical Database Systems, Privacy
Preserving Data Mining Techniques. At the University of Massachusetts-Lowell, Georges Grinstein is
teaching a new visual analytics course that he developed. The course
discusses and combines principles from cognitive science, information
visualization, geospatial information systems, amchine-based
rasoningand learning, and data mining to present a broad view of
visual analytics and its use by decision makers to detect the expected
and discover the unexpected. Students enrolled in the course are
gaining an understanding of the fundamentals of visual analytics and
its applications, an understanding of the analytical reasoning
process, an understanding of cognition, perception, and the ability to
design, build, and evaluate suitable visual representations of a
real-world dataset for decision makers of the public.
As noted in the discussion of Education Project 4, Steve Miller at
Williams College and Midge Cozzens at Rutgers University are working
on a book in Cryptography, with applications to Homeland
Security for undergraduate Liberal Arts and Criminal Justice
Students, who are likely to pursue careers in the FBI, CIA, Homeland
Security Agencies, etc. The book has now been pilot tested twice,
once in a course at Williams, a January term course for sophomores
through seniors not majoring in STEM disciplines, and once in a course
at Rutgers called Topics in Mathematics for Liberal Arts and aimed at
freshman through seniors not majoring in STEM disciplines. As we
continue to develop the book, we hope to make the courses
permanent. During this year, we also collaborated with our colleagues at the
Transportation Security COE to develop a Certificate program at
Rutgers in homeland security. The first such certificate emphasizes
transit security but utilizes some courses given by CCICADA faculty.
The goal is to modify this certificate program to give it several
different areas of emphasis, with the same overall structure. Interfaces with Research Tasks/Projects at CCICADA: The research task, such as optimal sampling (8.2), anonymization
(6.2), and differential privacy (6.3) have already given rise to new
courses. Connections to work at other COEs and VACCINE: We hope to develop interdisciplinary coursesthat will benefit from
advice of our partner COEs and VACCINE. Homeland Security Motivation New course and certificate programs are aimed at helping to develop
the homeland security workforce of the future Connection with Homeland Security Practitioners: As new courses are develped, we anticipate involving homeland
security practitioners as guest lecturers. Proposed Work in Year 2: We will continue to work on developing new courses and new curricula. CCICADA Education Project 10: Connections with ADMI (Association
of Computer/Information Sciences and Engineering Departments at
Minority Institutions) Project Leader: Project Participants: Project Description: We will support meetings and other efforts of the Association of
Computer/Information Sciences and Engineering Departments at Minority
Institutions (ADMI) and the Alliance for the Advancement of
African-American Researchers in Computing. it should be noted that
this was not included in the Year 1 workplan as it only arose late in
Year 1. While this was not given any specific funding, it presented
an excellent opportunity to collaborate with our partnerVACCINE, which
is in our workplan, and we used some of the CCICADA management travel
funds to send Jack Jarmon to participate this year. Accomplishments to Date: The Association of Computer/Information Sciences and Engineering Departments at Minority
Institutions (ADMI) and the Alliance for the Advancement of
African-American Researchers in Computing held their first combined
symposium to disseminate and discuss education and research
initiatives in the computing fields. The conference called "Winds
of Change in Computing" occurred between april 8-10, 2010, in Jackson,
Missippi. Representatives from private industry, universities,
foundations, federal government agencies, and national laboratories
addressed students and faculty on trends and opportuntities for
advancing computer scientists and engineers of underrepresented
populations. Speaking on behalf of the combined CCI Center, Marti Burns of
VACCINE and Jack Jarmon of CCICADA described the CCI's Center's
research role and educational mission. Participants were told that,
with the objective to help educate the next generation of Homeland
Security professionals, CCICADA offers Minority Serving Institutions
access to resources not traditionally available. Through CCICADA's
expansion of existing curricula, research seminars, and summer
research programs, MSIs are not only active participants, but also
part of the network of institutions and professionals working to
address the challlenges of managing and distributing actionable
information amid the current onslaught of unstructured data. These
programs span the entire educational pipeline. From K-12 through
graduate level work and professional training, VACCINE and CCICADA are
committed to building a foundation, which will serve the national
security needs of today and the future. In quest of that goal,
student and faculty attendees were invited to learn more. Work to be done the Rest of Year 1: We will continue the dialogue with ADMI and seek to find ways to
engage with it in the future> Interfaces with Research Tasks/Projects at CCICADA: Research at CCICADA informed the presentation made by Jarmon Connections to other work at COEs and VACCINE: This project was jointly develped with VACCINE. Homeland Security Motivation This project aligns with the DHS goal of developing the homeland
security workforce of the future. Proposed Work in Year 2: We plan to continue engagement with ADMI. Because the budget is so
tight, we do not propose a budget allocation for this purpose in the
first phase of funding for year 2. We will continue to use management
travel funds and the management discretionary fund as a ppropriate to
support this activity.
We have been building the DSSI program since its inception in 2007 and
each year more improvements are made to attract top Computer Science
students from around the U.S. The bulk of the year 2 effort will be
will be to plan for the summer 2011 DSSI. Plans for the 2001 DSSI
would be to increase the enrollment if possible, thus giving more
students from across the U.S. the opportunity to work with world class
researchers and to hear and network with industry leaders in the Data
Sciences. However, unless we receive more funds in a second
installment, we will have to cut the size of the program in Year
2.
Margaret Cozzens, PI, Rutgers University
Asamoah Nkwanta, Morgan State University
Abdul-Aziz Yakubu, Howard University
Georges Grinstein, UMass-Lowell (principal Lecturer in
2009)
John Stasko, Georgia Tech (principal Lecturer in
2009)
Eduard Hovy, USC (principal Lecturer in
2010)
Dan Roth, UIUC (principal Lecturer in
2010)
Tamara Carpenter,Rutgers University
Gene Fiorini,Rutgers University
Jack Jarmon,Rutgers University
Fred Roberts,Rutgers University
Heather Byrne, University of Massachusette Lowell
Curren Kelleher,University of Massachusette Lowel
Shawn Konecni,University of Massachusette Lowel
Reconnect Conference 2009:
Dates: August 2-15, 2009
Location: Rutgers University
Theme: Visual Analytics and its Applications
Mahmud Akelbek, Texas State University
Earl R. Barnes, Morgan State University
Loften Bullard, Florida Atlantic University
Joyati Debnath, Winona State University
Srabasti Dutta, college of Saint Elizabeth
James D. Factor, Alverno College
Maria Fung, worchester State College
Lila Ghermi, Texas Southern University
Craig Haile, college of the Ozrks
Andrew Jones, florida A&M University
David Miller, West Virginia University
Edmond Nadler, East Michigan University
Oscar Ortega, Harold Washington College
Michael Pelsmajer, Illinois Institute of Technology
Bala Ram, North Carolina A&T University
Ahlam Tannouri, Morgan State University
Sam Tannouri, Morgan State University
Ilya volnyansky, Canada Border Services Agency
Talitha M. Washington, University of Evansville
Reconnect Conferrnce 2010:
Dates: June 6-12, 2010
Location: University of Southern California
Theme Extracting and Visualizing Information from Natural
Language Text
course materials for the 2010 Reconnect will be developed andwe will
hold Reconnect 2010
We are working very closely with VACCINE to develop and run the
Reconnect conferences. The data sets used in the first Reconnect were
from the IEEE VAST Challenge (generated by PNNL and UMass-Lowell).
Our June 2010 reconnect will be hosted by our partner USC/ISI. The
DHS CREATE Center has been tremendously helpful in arranging for us to
use their faciities.
Midge Cozzens, Rutgers University
Fred Roberts, Rutgers University
Asamoah Nkwanta, Morgan State University
Abdul-Aziz Yakubu, Howard University
Plus many CCICADA members, especially Rutgers, but also at AT&T
labs and Bell Labs, who are acting as mentors for REU students.
The CCICADA research projects in REU are all chosen from the work that
either has direct connection to a homeland security
problem/application and aims to develop a methodology that is relevant
to or motivated by such a problem.
Margaret Cozzens, PI, Rutgers University
Center-wide participants from CCICADA and VACCINE
Tamra Carpenter, Rutgers University
Midge Cozzens, Rutgers University
Fred Roberts, Rutgers University
Date: November 16-18, 2009
Organizer: Nina Fefferman (Rutgers)
- To what extent can we combine information from the
different types of studies and how do we do it? Can we correct data
quality problems across multiple data sources?
- Can we use multiple data sources for improved reasoning
with uncertain data?
- How do we combine expert opinions or intelligence
information (possibly using machine learning approaches) with more
quantitative data?
- Should we integrate databases from different studies or
should we make combined inferences over different studies?
- How can we combine multiple views to rank items in
recommendation systems?
- Can we assure privacy with little or no negative impact on
the quality of the results?
Date: May 6-7 2010
Organizaers: Tamra Carpenter (Rutgers), Minge Xie (Rutgers)
Date: June 3-4, 2010
Organizers: Melike Baykal-Gursoy (Rutgers), Endre Boros (Rutger), and
Nina Fefferman (Rutgers)
Workshops and seminars are important vehicles for cross-fertilization,
dissemination of results, and infusion of new ideas into CCICADA
projects. As such, they are intended to complement and connect with
topics at the heart of CCICADA project research. For example, the
workshop on Mathematicl Models for Behavioral Epidemiology closely
relates to to research tasks 5.3 and 9.1. The workshop on statistical
Issues in Analyzing Information on Diverse Sources addresses a topic
that is central to many CCICADA projects, particularly those
represented in tasks 3.2, 3.3, and 3.4. The proposed year 2 workshop
on Adversarial Decision Making strongly relates to tasks 4.3 and
9.3.
Date: September 30t -October 1, 2010
Organizers: David Banks (Duke), Janusz Marecki (IBM T.J. Watson
Research), Bonnie Ray (IBM T.J. Watson Research), Milind Tambe
(CREATE/ University of Southern California)
Midge Cozzens, Rutgers University
Fred Roberts, Rutgers University
Asamoah Nkwanta, Morgan State University
Abdul-Aziz Yakubu Howard University
- Professional/Career Development
- Scholarly Productuvity
- Student Development
- Databases
DATE: November 12-14, 2009
LOCATION: The University of the District of Columbia
Dawn Lott, National Association of Mathematicicans (NAM)
Leon Woodson, National Association of Mathematicicans (NAM)
NAM MATHFest is directly in line with the DHS goals of preparing the
homeland security workforce of the future.
Nancy Komlan, University of Illinois at Urbana-Champaign
Dan Roth, University of Illinois at Urbana-Champaign
George Reese, University of Illinois at Urbana-Champaign
Kathleen Harness, University of Illinois at Urbana-Champaign &
the Dr. Howard Elementary School
Daniel Wendel, StarLogo at Massachusetts Iinstitute of Technology
Other MSTE and MIT staff will be joining in the coming year as we
complete the full proposal
Midge Cozzens, Rutgers University
Fred Roberts, Rutgers University
Warren Powell, Princeton University
Lila Ghemri, Texas Southern University
Georges Grinstein, UMass-Lowell
Midge Cozzens:, Rutgers University
Marti Burns, Purdue University
Jack Jarmon, Rutgers University