|
Benchmark Intelligent Agent Systems for Distributed Battle Tracking
Dr. Albert Esterline
We are developing benchmark intelligent agent systems for the US Army CERDEC C2D (Command and Control Directorate) as an integral part of this directorate’s development of distributed battle command tracking technology. Initially, the focus is on developing a benchmark in the JADE (Java Agent DEvelopment) framework, a software framework implemented in Java that simplifies implementation of multi-agent systems through a middle-ware compliant with the FIPA specifications. Other agent frameworks may be investigated later in the project. Web services are being incorporated into the systems using the WSIG add-on to JADE, and the Jess rule engine is being used to implement decision rules. This effort will contribute directly to CERDEC’s mission, “To develop and integrate Command, Control, Communications, Computers, Intelligence, Surveillance, and Reconnaissance (C4ISR) technologies that enable information dominance and decisive lethality for the networked Warfighter”.
NOAA ISET Cooperative Research and Education Center, Research Area III-C (Distributed Architectures)
Dr. Albert Esterline, Dr. Yaohang Li,
Dr. Matt Heavner (Dept. of Environmental Science, Univ. of Alaska Southeast)
Grid-computing and multiagent techniques are being developed to support data acquisition, fusion, and mining in collaboration with the NOAA Earth System Research Laboratory (ESRL). We are applying grid computing for real-time data management in large sensor networks. Multivariate stream data mining is used to detect spatio-temporal patterns spanning multiple sensor streams. Multiagent protocols are being used to coordinate the wide range of sensors, analysis techniques, and subsystems required to identify complex, dynamic relationships in large volumes of data. In light of the problems with testing, we model-check critical aspects of the concurrency of the multiagent architectures developed. Initial deployment is with the wireless sensor web for glacial monitoring maintained by the University of Alaska Southeast.
AttackAids@A&T
Dr. Gerry Dozier, Dr. Kelvin Bryant, Jeremy Barksdale, Gerrimy Tann, Toyan Harvey, Marcel Brown, Lamar Simpson, Robert Meeks, Dominique Hairston, Karlen Marshall, Tim Watson, Sir Harvey
The AttackAids@A&T Project is a NSF sponsored project (via the Broadening Participation in Computing program) and is part of the Alliance for Advancing African-American Research in Computing (A4RC). In this project, graduate as well as undergraduate students are collaborating with researchers from the University of Missouri, Virginia Tech, Auburn, and NASA in developing High-Performance Genetic & Evolutionary Computations (GECs) for solving nk-HIV Protease Docking Problems (nk-HPDPs).
The AttackAids@A&T Project is currently devoted towards discovering the underlying properties associated with efficient heterogeneous distributed GECs and how theyshould be assembled. We are currently using the GECs found in X-TOOLSS (eXploratory Toolset for the Optimization Of Launch and Space Systems) (http://www.eng.ncat.edu/dept/cs/ACIwebPage/xtoolss/index.htm) developed by the Applied Computational Intelligence Lab.
A TASK for Being Me, Myself, and I
Dominique Hairston, Darliene Hopes, Tamyrah Mack, Richard Messick,
Nakia Powell, Donovan Reese, Dr. Gerry Dozier
One of the biggest threats to cyber security is unauthorized access via social engineering. Through this method hackers gain access to databases and files that contain sensitive information such as usernames and passwords. With this information in hand, hackers may then gain easy access to computer systems. In this paper, we present TASK (Tiny Authentication Security Kit), a Kohonen-based anomaly detection system developed as a means to combat this threat. TASK is a system that interacts with the user to build a profile and then uses this profile to build a login signature for the purpose of anomaly detection. TASK consists of three components: (1) an anomaly detection system for correct logins of an authorized user, (2) an anomaly detection systems for incorrect logins of an authorized user, and (3) an interactive evolutionary-based vulnerability analyzer for discovering potential holes in both systems before hackers have a chance to find them. In this research project, we also compare how the layout of the TASK interface (the location of the login button, username & password boxes, and the submit button) impacts the accuracy of the underlying anomaly detection system.
An IDEA for Computational Hip-Hop
Dr. Gerry Dozier, Jeremy Barksdale, Dr. Kelvin Bryant, Lamar Simpson,
Kevin Moore, Zelanie Johnson, Marcus Hicks
The field of Evolutionary Computation is devoted towards the study, analysis, and design of problem solvers based on simulated evolution. Evolutionary Computations (ECs) are quite different than other traditional search methods in that they evolve a population of candidate solutions to a problem rather than working with a single candidate solution. Perhaps the most important component of an EC is its evaluation function. The evaluation function is used to assign candidate solutions a fitness (a relative measure of ‘goodness’). The EC then selects parents to create offspring based on their fitness. The better a candidate solution’s fitness the more opportunities it has to procreate.
For some problems, it is difficult to develop an evaluation function in the form of a closed form mathematical equation to be optimized. For these types of problems, interactive evolutionary algorithms are a welcomed alternative. In an interactive evolutionary algorithm, the evaluation function is replaced by a user. The user evaluates candidate solutions and provides each with a user specified fitness. Thus, it is the human user that guides the evolutionary process.
In this research project, we will present an interactive distributed evolutionary algorithm (IDEA) for evolving the play sequence for an eight song Hip-Hop CD. The IDEA will allow multiple users to interactively evolve play sequences concurrently. Our results show that play sequences evolved by a group of individuals working together (via IDEA) have greater appeal to listeners than play sequences interactively evolved by an individual.
Tutoring to Apply Conceptual Understanding of Ordinary Differential Equations: Analysis for Cognitive and Computable Patterns
Dr. Jung Hee Kim
We have been engaged in analyzing tutorial dialogs for intelligent tutoring system construction. Our Wooz-Tutor project enables mathematics tutoring, where the tutor and student communicate through a computer. The program provides optional assistance to the tutor. The Wooz partly-automated model should allow us to build and test incomplete and imperfect software components. From January 2007 we have started working on tutoring differential equations for conceptual understanding, a relatively new area of mathematics education research. This research is funded by NSF.
Advanced Sensor Network Applications for Environmental Monitoring Systems
Dr. Jung Hee Kim and Dr. Jung Hyeon Kim (EE dept. at NCA&T)
We have been working on the Advanced Sensor Network project, an NSF-funded interdisciplinary collaboration between the Electrical Engineering, Natural Resources and Environmental Design, and CS departments at NCA&T. Dr. Jung Hee Kim’s research involves automatic coding of sensor network data for data mining. We are creating an automatic coding mechanism robust enough to operate in the Advanced Sensor Network’s dynamically changing environment.
Collaboration through Agile Software Development Practice
Dr. Jung Hee Kim, Dr. Kelvin Bryant and Jeremy Barksdale
This is a collaborative project spearheaded by UNCC that seeks to measure the efficacy of agile development practices, particularly peer programming, in upper class programming courses. In this grant, we focus on the potential for agile practices to improve the participation and success of various underrepresented groups of the IT workforce. Students will be paired and assigned a software engineering project to be completed over the semester. They will be instructed on Pair Programming techniques and encouraged to use these techniques in the completion of their project. We will gather information from their experiences and will compare results from students who did and did not use agile methods.
High Resolution ab initio Protein Folding
Dr. Yaohang Li
The successful ab initio protein structure prediction depends on the surmounting of three efforts: (1) formulating an accurate and sensitive scoring function that can lead the search process to the global minimum in the protein folding energy landscape; (2) devising efficient moves (conformation changes) toward the native conformation; and (3) developing a global optimization algorithm that can efficiently escape from the deep local minima and converge to the global energy minimum. Among these three efforts, building an accurate and sensitive scoring function is of the most importance. However, just like many other computational biology problems, developing a sensitive and accurate scoring function is a very difficult and even formidable job. In reality, even though many scoring functions based on various criteria, such as energy, statistics, secondary structure, loops, or contact pairs, have been proposed, currently there does not exist a reliable and general scoring function that can always drive a search to a native fold, and there is no reliable and general global optimization method under these scoring functions that can sample the conformation space adequately to guarantee a significant fraction o near-natives (<3.0 A RMSD from the experimental structure). We seek to develop novel Monte Carlo approach to address the issue of the insensitivity in the current existing protein folding scoring functions and perform convergence analysis of the Monte Carlo sampling process. We are expected to develop protein modeling software tools to improve the resolution of ab initio protein structure prediction.
Grid-based Monte Carlo and Quasi-Monte Carlo
Dr. Yaohang Li
Monte Carlo applications are widely perceived as computationally intensive but naturally parallel. Therefore, they can be effectively executed on the grid using the dynamic bag-of-work model. We improve the efficiency of the subtask-scheduling scheme by using an N-out-of-M strategy, and develop a Monte Carlo-specific lightweight checkpoint technique, which leads to a performance improvement for Monte Carlo grid computing. Also, we enhance the trustworthiness of Monte Carlo grid-computing applications by utilizing the statistical nature of Monte Carlo and by cryptographically validating intermediate results utilizing the random number generator already in use in the Monte Carlo application. All these techniques lead to a high-performance grid-computing infrastructure that is capable of providing trustworthy Monte Carlo computation services. These techniques can be also extended to quasi-Monte Carlo applications.
Developing Interactive Tools for Firewall Training
Dr. Ken Williams
Firewalls are used as a perimeter defense against network intrusion. The configuration of firewalls may be confusing and subtle. We are developing an interactive “game like” training system to help teach the effective configuration of firewalls. In the distributed training system each student has a simulated computer system they are responsible for defending. They can simultaneously launch attacks against the systems of other students while configuring their firewall to defend against similar attacks. A simulation manager occasionally provides new requirements for the student’s simulated systems while scoring each student’s successes.
Open Source Reconfigurable Computing Framework
Dr. Ken Williams and Dr. Christopher Doss
Reconfigurable Computing is a field of computing that combines a general purpose processor with a Field Programmable Gate Array (FPGA). The use of FPGA’s offers the ability to create applications that are significantly faster than the corresponding implementation on a typical general purpose processor. However, the potential performance offered by these devices is hard to exploit due to the difficulty of mapping an algorithm to the FPGA.
The purpose of the proposed project is the development of tools (framework) that facilitate the pooling of resources that ease the design and implementation of software applications on reconfigurable computing systems. We propose to develop this framework by leveraging the use of the commonly used Eclipse Platform. By leveraging the Eclipse Platform, especially its use of plug-ins, we anticipate that many researchers will be able to add their expertise to the development of a refined development platform, thus broadening the impact of reconfigurable computing in general.
Detecting and Responding Network-centric Attacks
through Intelligent Visual Analysis
Dr. Huiming Anna Yu
This exploratory research is to build highly scalable and interactive solutions to prevent and respond to network-centric attacks. In this project we will design and implement an interactive real-time visualization system to monitor large scale-scale network traffic activity, to analyze the large amount of data, to identify abnormal behavior, and to proactively respond (manual or automatic) to network-centric attacks; and c) provide an adaptable and expandable environment to meet the needs of the administrators or researchers.
A Hybrid Detection, Analysis and Response Method for DNS Amplification Attack
Dr. Huiming Anna Yu
Domain Name Service (DNS) attack is a new form of denial of service attack based on the difference in size between a Domain Name System query and a DNS response and the willingness of DNS servers to answer queries from any source. Attackers exploit recursive name servers to amplify Distributed Denial of Service (DDoS) attacks by utilizing IP spoofing. In this project we will develop a hybrid method that will integrate agent technology, visual analytic and interactive visualization techniques to allow system administrators to interact with the system in real-time, to visualize network traffic, to analyze the collected data and to detect the DNS Amplification Attack.
Developing Visualization Tools for Teaching Information Security
Dr. Xiaohong Yuan
In this project, we have developed interactive visualization tools for packet sniffer, Kerberos authentication architecture, and wireless LAN attacks. These tools have been used in teaching computer networks and security courses to demonstrate information security related concepts. Currently, we are developing more visualization tools for demonstrating wireless network attacks. We are also developing hands-on labs for using open-source tools to conduct wireless network penetration testing. The visualization tools we developed can be accessed at: http://www.ncat.edu/~xhyuan/security_visual_tools/
Integrating Software System Security Evaluation into Computer Science Curriculum
Dr. Xiaohong Yuan and Dr. Yaohang Li
The goal of the project is to integrate secure software system evaluation into Computer Science curriculum by developing course modules for “Operating Systems”, “Database Design” and “Software Engineering”. The software engineering course module can include subjects such as penetration analysis, formal method, fault injection methods, developing risk-based test plan using risk analysis and threat modeling, etc. The database course module can include threat modeling of database applications, vulnerabilities such as database overflow, SQL Injection, etc. The operating systems course module can include operating system related vulnerabilities such as buffer overflow with environment variables, API calls and local command line utilities, stack overflow, heap overflow, etc.
Developing Case Studies for Information Assurance Education
Dr. Xiaohong Yuan, Dr. Huiming Anna Yu
The goal of this project is to develop new case study course materials and methods to support the wide adoption of the case study teaching approach in undergraduate IA education and to enhance IA education nationwide. We plan to develop case studies in four areas: (1) network security; (2) security management; (3) web security; (4) application security. These cases will incorporate hands-on laboratory components, such as open source tools, interactive simulations, cyber games, visual aids and multi-media presentations. The developed cases will be used in the existing information security undergraduate and graduate courses currently offered at NC A&T SU.
Data mining of relationships between keywords in social tagging website
Dr. Jinsheng Xu
Social tagging website like del.icio.us lets users to classify their bookmarks through “tagging”. The users choose one or more keywords to give some “meaning” to their bookmarks. There are many-to-many relationships among users, keywords and bookmarks. This project explores some “interesting” relationships between keywords. E.g. are there “is-a” and “also interested” relations among keywords?
Understanding how collaborative knowledge process of Wikipedia works with agent-based simulation
Dr. Jinsheng Xu
The knowledge in Wikipedia is produced and processed collaboratively by online user community. The results of this collaboration process present various seemingly complex patterns demonstrated by update history of different articles in Wikipedia. Agent simulation is a powerful method that is used to study the behaviors of complex systems of interacting and autonomous agents. In this project, we study the collaborative knowledge processing in Wikipedia using a simple agent-based simulation model.
Analyzing the update history of Wikipedia articles for trustworthy analysis and vandalism detection
Dr. Jinsheng Xu
Wikipedia articles can be arbitrarily modified by virtually any user. Some articles in Wikipedia are heavily vandalized and many current articles may contain incorrect information. However, previous studies of Wikipedia show that false knowledge is reverted or undone quickly by the large user community of Wikipedia. Part of an article that has long lifetime and have endured many updates tends to be more trustworthy than the parts that are updated recently. In this project we study update history of Wikipedia articles to find characteristics of vandalisms.
Hamiltonain Cycles in “near” Cayley color digraphs
Professor Edward C. Carr and Dr. Joseph B. Klerlein (Western Carolina University)
Through the use of computation we are discovering Hamiltonian Cycles in vertex transitive digraphs and “near” Cayley color digraphs. We are actively developing algorithms for finding Hamiltonian cycles in these graphs and proving the hamiltonicity of these graphs. Our referred publications on this subject have been and continued to be published in the Graph Theory Journal Congressus Numerantium.
|
