Main Page
From CIMIC Events
Contents |
CIMIC Rutgers: Activities and Events | CIMIC NEWS & Announcements
October 9 2009
Dissertation Proposal-- Haibing Lu
Boolean Matrix Decomposition with Extensions: Application to Role Mining
October 9 2009
11:00 am
RBS 1198
Abstract
Boolean matrix decomposition (BMD) refers to decomposing of an input Boolean matrix into a product of two Boolean matrices, where the first matrix represents a set of meaningful concepts, and the second describes how the observed data can be expressed as combinations of those concepts. As opposed to standard matrix factorization, BMD focuses on Boolean data and employs Boolean matrix product instead of standard matrix product. The key advantage of BMD is that BMD solutions provide much more interpretability, which enable BMD to have wide applications in multiple domains, including role mining, text mining, feature extraction, dimensionality reduction, information retrieval, clustering, and data compression.
Real applications carry varying expectations on BMD solutions, and various constraints associated with concrete circumstances also apply. Incorporating those objectives and constraints makes the task of searching for a good BMD solution nontrivial. Apart from that, BMD by itself has the issue of insufficiency in modeling some real data semantics, as only the set union operation is employed in combination. This way of modeling incompletely represents some real data semantics, and hence inhibits the interpretability of the BMD solutions.
Hence, the objective of this dissertation is two-fold: (i) First, it builds a unified framework to encompass all BMD variants. To do so, we will review real applications of BMD and categorize them by extracting representative BMD variants. Then we will formulate those BMD variants through Integer Programming (IP). Those IP formulations will allow us to directly adopt good IP software packages and algorithms. Concerned with scalability, we will study computational complexity for those BMD variants. To make those NP-hard BMD variants applicable, we will design efficient heuristics at the cost of losing accuracy. (ii) Second, it addresses the insufficiency of BMD in modeling. We propose a new notion, extended Boolean matrix decomposition (EBMD), which allows a data record to be expressed as an inclusion of a subset of concepts with an exclusion of another subset of concepts. Besides formulating representative EBMD variants through IP, we will study their computational complexity and propose efficient heuristics for large scale problems. These two objectives are pursued in the context of the role mining problem
April 3 2009
Invited Talk-Dr. Shamik Sural
Detection of Intrusive Activity in Databases by Combining Multiple Evidences and Belief Update
April 3, 2009
10:30 am
MEC 309
Abstract
We propose an innovative approach for database intrusion detection which combines evidences from current as well as past behavior of users. It consists of four components, namely, rule-based component, belief combination component, security sensitive history database component and Bayesian learning component. The rule-based component consists of a set of well-defined rules that give independent evidences about a transaction's behavior. Dempster-Shafer's theory is used to combine multiple such evidences and an initial belief is computed. First level inferences are made about the transaction depending on this initial belief. Once the transaction is found to be suspicious, belief is updated according to its similarity with malicious or genuine transaction history using Bayesian learning. Experimental evaluation shows that the proposed intrusion detection system can effectively detect intrusive attacks in databases without raising too many false alarms.
Bio
Shamik Sural is an Associate Professor at the School of Information Technology, IIT Kharagpur India. He received the B.E. degree in Electronics & Tele-communication Engineering from Jadavpur University, Calcutta, India, in 1990, M.E. in Electrical Communication Engineering from Indian Institute of Science, Bangalore, India, in 1992 and the Ph.D. degree from Jadavpur University in 2000. Before joining IIT, he held technical and managerial positions in a number of organizations both in India as well as the USA. Dr. Sural's research work has been funded by the Ministry of Communication and Information Technology, Department of Science and Technology, Govt. of India and the National Semiconductor Corporation, USA. His research interests include database security, data mining and multimedia database systems.
March 11 2009
Dissertation Proposal
Correlation Computing in Dynamic and Complex Data
Wenjun Zhou
March 11, 2009
11 am
MEC 309
Abstract
Correlation computing refers to the problem of efficiently finding groups of strongly related data objects in very large databases. Most previous studies have been focused on static data sets. However, in real-world applications, input data are often dynamic and must continually be updated. With such large and growing data sets, new research efforts are expected to develop an incremental solution for correlation computing. Along this line, we propose a CHECK-POINT algorithm that can efficiently incorporate new transactions for correlation computing as they become available. Specifically, we set a checkpoint to establish a computation buffer, which can help us determine an upper bound for the correlation. This checkpoint bound can be exploited to identify a list of candidate pairs, which will be maintained and computed for correlations as new transactions are added into the database. However, if the total number of new transactions is beyond the buffer size, a new upper bound is computed by the new checkpoint and a new list of candidate pairs is identified. Experimental results on real-world data sets show that CHECK-POINT can significantly reduce the correlation computing cost in dynamic data sets and has the advantage of compacting the use of memory space.
In this dissertation, we also propose to deal with correlation computing in data with complex data types. There are several research issues to be addressed. First, it is necessary to have new definitions of correlation patterns, such as local association and clique-type association types, for multi-level and multi-scale data. Second, it is interesting to develop methods for measuring associations among complex data, such as sequences and graphs. Towards this direction, we propose to design a directionbased correlation computing framework which can be applicable for characterizing and understanding the general movement patterns and detect outliers in location traces. Finally, we will study parameter estimation based on pattern preserving sampling, which can help users choose suitable parameters for correlation computing.
Bio
Wenjun Zhou is currently a PhD student at the Department of Management Science and Information Systems at Rutgers Business School. She received her B.S. degree in Information Management and Information System from the University of Science and Technology of China - Hefei, China and a M.S. degree in Biostatistics at the University of Michigan - Ann Arbor. Her research interests include data mining, statistical computing, and their applications in business and management science.
March 6 2009
Doing Business with DHS
See this agenda in printer friendly PDF format [1].
Location: Center for Law and Justice (Law School) Room- 090
123 Washington Street
Newark, NJ 07102
973-353-1014/ 5561
March 6, 2009
12:30 pm – 5:00 pm
10:00- 10:30
Meeting with:
Thomas Cellucci, Chief Commercialization Officer of the Science & Technology Directorate, DHS
Nabil Adam, Professor and Director, Rutgers CIMIC
CIMIC Faculty and PhD. Students
10:30- 11:00
Meeting with:
Michael Cooper, Dean of Rutgers Business School
Thomas Cellucci, Chief Commercialization Officer of the Science & Technology Directorate, DHS
Nabil Adam, Professor and Director, Rutgers CIMIC
11:30-12:30
Meeting with:
Steven Diner, Chancellor, Rutgers Newark
Thomas Cellucci, Chief Commercialization officer of the Science & Technology Directorate, DHS
Nabil Adam, Professor and Director, Rutgers CIMIC
12:30- 1:00
Lunch (Start of meeting- Law School- 090)
1:00- 1:15
Introduction, Nabil Adam
1:15- 2:15
Presentation, Thomas Cellucci
2:15- 2:30
Break
2:30- 5:00
Open discussion: Dr. Thomas Cellucci, Corporate Guests, Faculty, and Students
February 6 2009
MEC 215
11:00 - 1:00
'Invited Talk-'Prof. Mira Balaban
Title: Correctness of UML Class Diagrams
UML is now widely accepted as the standard modeling language for software systems. It plays a central role in the emerging model driven engineering approach for software development. Therefore, it is essential that UML modeling tools are strengthened with features for identification of erroneous models, detection of sources of errors, revealing redundancies, and providing design improvement advice. Early problem detection might prevent major future faults.
The Class Diagram is UML core view, having well formed semantics and providing the backbone for any modeling effort. Class diagrams are widely used for purposes such as software specification, database and ontology engineering, meta-modeling and model transformation. Correctness of UML class diagrams refers to the capability of a diagram to denote a finite but not empty reality. This is a natural, unquestionable requirement. Nevertheless, incorrect diagrams are often designed, due to interaction of contradicting constraints, and current CASE tools fail to detect erroneous designs.
In this talk I demonstrate correctness problems, caused by interaction of class diagram constraints, and describe state of the art techniques for identifying class diagram correctness problems, and detecting their causes.
Bio: Mira Balaban is affiliated with the Computer Science Department
in Ben-Gurion University in ISRAEL. She holds a Ph.D. in Computer
Science from the Weizmann institute, and a Music diploma in Piano
Performance from the Rubin Music Academy in Tel-Aviv University in
ISRAEL. Her research interests are in the areas of Software
Engineering, Conceptual Modeling, Knowledge representation, Database
Semantics, Programming Languages, and Computer music.
December 10 2008
Homeland Security Day at Rutgers CIMIC
Center for Law and Justice, Room B025
123 Washington Street
Newark, NJ 07102
973-353-1014
December 10, 2008
11:30 am – 3:15 pm
11:00- 12:00 Meeting with:
Steven Diner, Chancellor, Rutgers Newark
Philip Yeagle, Dean of Faculty of Arts & Sciences, Rutgers Newark
Alexander Gates, Professor and Chair, Earth & Environmental Sciences Dept.
Nabil Adam, Professor and Director, Rutgers CIMIC
Christopher Doyle, Director, Infrastructure & Geophysical Division, DHS-S&T
Lawrence Skelly, Deputy Director, Infrastructure & Geophysical Division, DHS S&T
11:30- 12:40 Working Lunch
12:00- 12:30 Introduction, Infrastructure & Geophysical Division, DHS-S&T
12:30- 12:40 Remarks, Michael Cooper, Dean, Rutgers Business School
12:40- 12:55 Overview of CIMIC Research Work, Nabil Adam
12:55- 1:30 Demo, iTEAM, CIMIC
1:30- 2:20 Individual CIMIC Research Presentations:
Vijay Atluri
Jaideep Vaidya
Basit Shafiq
Soon Ae Chun
Hui Xiong
Alex Borgida
S. Muthukrishnan
2:20- 3:15 Discussion:
Cherrie Black, NJ OHSP
John Dalton, CACI
John Ellenberger, SAP Research
Beverly Hawkswell, Eugene Olsen, John Keenan, ARDEC
Mohsen A. Jafari, Yan Lu, Dong Wei, John Pearson, Siemens
Richard W. Kelly, Director, ROIC-NJSP
John Kurmer, John Thoma, Battelle
November 19 2008
Dissertation Proposal
Security and Privacy in Personalized Mobile Service Environments
Heechang Shin
November 19, 2008, 1:00pm
MEC 215, Newark Campus
Dissertation Chairs: Dr. Vijay Atluri and Dr. Jaideep Vaidya
Services in a mobile environment are based on the locations of mobile users. Personalization, based on the profiles of mobile users, significantly increases the value of such services. However, they pose significant security and privacy challenges; Ensuring security and privacy for a personalized mobile environment in an efficient manner is the primary objective of this dissertation. To this end, we will address the following research issues. Often, access control requirements in a mobile environment are based on the spatio-temporal attributes of mobile users, resources to be protected, profiles of users, or all of these. Evaluating an access request requires searching for the desired moving objects that satisfy the query as well as identifying and enforcing the relevant security policies. Enforcing these access control policies incurs significantly more overhead than that in a traditional environment as the security policies need to be searched based on space and time rather than the user or object identifiers. It has become the norm to
index spatio-temporal objects for improving the query performance. Employing similar index structures for security policies may result in efficient enforcement. Instead of having a separate index for mobile objects, security policies and profiles, this dissertation will develop a unified index structure capable of indexing all three in a single index. We plan to conduct a comprehensive experimental evaluation to examine its scalability and performance.
Another pressing issues in delivering mobile services is in protecting the privacy of users. The concept of location $k$-anonymity has been advanced to address mobile user privacy. However, it is not sufficient to comprehensively protect privacy in the personalized mobile service environment due to the additional background knowledge such as profile and movement information that can be exploited by the adversary. In this dissertation, we will also propose a more comprehensive family of anonymity models that incorporate location, direction, as well as profile information. We will propose algorithms that can constrain both the generalization of the location as well as that of profiles, while meeting the quality of service requirements. In addition, ensuring such anonymity can limit tracking of the service requestor while continuously receiving a service.
November 14 2008
Invited Talk- Prof Wendy Hui Wang- Steven Institute
ABSTRACT
Title: Privacy Preserving Data Sharing for Distributed and Heterogeneous Data
There is a growing demand for sharing data repositories that often
contain personal information across multiple autonomous and
heterogeneous databases. Such data sharing is subject to constraints
imposed by privacy of individuals or data subjects as well as privacy of
data providers. Concretely, we want to support queries across the
distributed and heterogeneous databases, but the data need to be
anonymized such that query results do not contain individually
identifiable information and data providers do not reveal their
databases to each other apart from the query results. This talk
discusses the challenges that arise in this context and presents a set
of initial solutions including: an integrated framework for anonymizing
both structured and unstructured data, a set of decentralized protocols
that allow independent data providers to build a virtual anonymized
database, and a distributed querying infrastructure that allow users to
query the virtual database.
November 4 2008
Invited Talk- Li Xiong,Assistant Professor of Computer Science at Emory University
ABSTRACT
There is a growing demand for sharing data repositories that often contain personal information across multiple autonomous and heterogeneous databases. Such data sharing is subject to constraints imposed by privacy of individuals or data subjects as well as privacy of data providers. Concretely, we want to support queries across the distributed and heterogeneous databases, but the data need to be anonymized such that query results do not contain individually identifiable information and data providers do not reveal their databases to each other apart from the query results. This talk discusses the challenges that arise in this context and presents a set of initial solutions including: an integrated framework for anonymizing both structured and unstructured data, a set of decentralized protocols that allow independent data providers to build a virtual anonymized database, and a distributed querying infrastructure that allow users to query the virtual database.
SPEAKER'S BIOGRAPHY
Li Xiong (http://www.mathcs.emory.edu/~lxiong) is an Assistant Professor of Computer Science at Emory University. She holds a PhD from Georgia Institute of Technology and an MS from Johns Hopkins, both in Computer Science. She also worked as a software engineer in IT industry for several years prior to pursuing her doctorate. Her areas of research are in distributed data management, privacy-preserving and trustworthy data sharing, and bio and health informatics.
September 24 2008
Invited Talk- Aunshul Rege, PhD. Rutgers School of Criminal Justice
ABSTRACT
Title Cybercriminology: Moving Towards a Theory for Crime and Criminality in Cyberspace
Computers and technology have impacted our society tremendously. The fusion of computing and communications has resulted in the creation of cyberspace, which transcends the physical domain. Digital environments and information and communications technologies (ICTs) have altered the organizational dynamics of crime, criminality, and criminal communication. This paper develops an integrated theoretical framework that operationalizes Deleuze and Guattari's rhizomatic model to account for the digital environment where cybercrimes occur, and applies Best and Luckenbill's theory on criminal organization to understand the scope and complexity of crimes in cyberspace. It accounts for how criminals organize, operate, and network in digital environments around the six dimensions of space, time, movement, scope, structure, and preservation. I argue that while my integrated theoretical framework adequately captures crimes and criminality occurring exclusively in cyberspace, it requires modification along the concepts of space, time, and structure to address the organizational traits of crimes and criminals in ‘hybrid’ space. I then evaluate my revised integrated theory along the criteria of scope, coherence, causality, and predictive power to demonstrate theoretical growth. Finally, I offer a theoretical proposition that can be used as a point of departure for future cybercrime studies in the criminological discipline.
SPEAKER'S BIOGRAPHY
Aunshul Rege is a PhD student at the Rutgers School of Criminal Justice. Her research interests include the organization of crimes and criminality in digital environments, and their implications for policing agents, the law and criminal justice system, and international policy. She is currently working on a thesis about cyberattacks against critical infrastructures.
August 19 2008
Invited Talk- Prof. Hulya Yildirim, Kocaeli University, Turkey
ABSTRACT
A review of satellite remote sensing studies in Turkey between 1985 to present, based on my personal experience will be presented.
As a developing country with relatively large surface area, Turkey is an ideal place to show the benefits of satellite remote sensing (SRS) and related information technologies such as Geographic Information Systems (GIS). Still it had its own difficulties and shortcomings. SRS had started as a research topic in a small group at a government laboratory in 1977. Later a larger group with an image processing laboratory (under the umbrella of Space Technologies Department) with more directed education and research program was established. Several courses to public bodies to disseminate the technology with application examples were prepared by our group. A UN - UNDP project to support the laboratory and educational activities were also undertaken.
I will also give examples of application projects undertaken by our group, mainly in the period between 1993 and 2005. Our applications were mostly on the monitoring of uncontrolled industrialization and fast urbanization (short and long term monitoring), assessment and impact of soil erosion, earthquake damage assessment during and after 1999 Marmara Earthquake, forest fires and their effects on the environment, watershed rehabilitation monitoring by SRS. As one can see, they are mostly agricultural and environmental in nature, with land-cover/land-use classification, their assessment and disaster related applications, mostly used through GIS, GPS and some other information technologies.
Later, I transferred from research center to university system working at, first, at Canakkale University, later at Kocaeli University. I started to share my experiences on local problems with my thesis students. Currently, I am serving as the head of Computer Engineering Dept of Kocaeli University. Our program includes SRS and GIS as one of the optional research and application areas.
SPEAKER'S BIOGRAPHY
Hulya Yildirim is currently (since 2006) working as a professor at the Computer Engineering Department of Kocaeli University, at Kocaeli, Turkey. She is also serving as the head of the same department. From 2002 to 2006, she was a professor at Computer Engineering Department at Canakkale University (COMU) at Canakkale, Turkey. Prof. Yildirim is the founder of Remote Sensing Research Center of COMU. Earlier to this date, she was a researcher at Marmara Research Center of Turkish National Research Council (TUBITAK) from 1978 to 2002. She received her PhD from Cukurova University, Department of Soil Sciences. Her PhD thesis was on the development, application and comparison of various classification methods under various conditions interest to agriculturists. She received a BSc degree in Mathematics from Middle East Technical University, Ankara, Turkey.
Her research interest in satellite remote sensing has been the comparative studies of various classification schemas, applications of land-cover/land-use classifications and also, the long term monitoring and assessment for a number of diverse agricultural and environmental problems in Turkey.
Getting started
- Consult the User's Guide for information on using the wiki software.
