Student Research Opportunities in the Department of Computer Science

• Some examples of graduate student research can be seen at the:
  
  • Fourth Annual Graduate Student Research Day
  • Fifth Annual Graduate Student Research Day
  • Sixth Annual Graduate Student Research Day
  • Seventh Annual Graduate Student Research Day
  • Eighth Annual Graduate Student Research Day
  • Ninth Annual Graduate Student Research Day
• Listings of past undergraduate research projects can be found at the Research Fellowship Program Site.

Project Description: A firewall is a system or group of systems that enforces a security policy between networks. Inspecting traffic sent between the networks, the firewall provides access control, auditing, and traffic control based on a security policy. It is important that the firewall acts transparently to legitimate users, with little or no effect on the perceived network performance. This is especially true if traffic requires specific network Quality of Service (QoS). The firewall should process the legitimate traffic quickly and efficiently. Unfortunately, the firewall can quickly become a bottleneck given increasing traffic loads and network speeds. This project will investigate new firewall architectures that is suitable for high speed networks and scalable for large traffic loads, maintain QoS requirements across the network boundary, and greatly lessen, if not eliminate, DoS attacks.

Project Title: QoS and Resource Management for Computer Networks
Faculty Director: Fulp
Subject Area: Network Management/Traffic Engineering
Level: Graduate Students

Project Description: Many network applications require minimum guarantees of network performance for their proper operation, such as bounds on the packet loss, delay, and jitter (varying packet interarrival time). These Quality of Service (QoS) guarantees can be provided with the proper allocation of network resources, such as link bandwidth, processor time, and buffer space. Unfortunately, users frequently contend for the finite resources that are available. As a result, most network QoS problems are caused by improper Resource Management (RM). There are several interesting RM problems associated with computer networks, consider the following list of topics:

Project Title: Wireless Network Testbed
Faculty Director: Fulp
Subject Area: Networks
Level: Undergraduate Students

Project Description: The Computer Science Department at Wake Forest University currently has a computer network laboratory, where research ideas can be implemented and tested under varying conditions. Currently the laboratory consists of configurable Linux-based routers and commercial wired network components; however, it relies on "wired connections" (Ethernet). This project will investigate the integration of wireless network elements and the development of future wireless networking course. Other laboratory projects include network security, and network simulators.

Project Title:  Genetic Algorithms for Computationally Intractable Problems  
Faculty Director: John
Subject Area: Genetic Algorithms
Level:  Undergraduate and Graduate Students

Project Description: A genetic algorithm is a heuristic for searching motivated by natural evolution.  This method is applied to problems that are at least NP-hard.  There is no guarantee that an optimal member of the search space will be found; but, the overall goal is to find very good members in polynomial time. I am interested in finding useful genetic algorithms for discrete problems, e.g. scheduling, language recognition, program generation, graph identification.  My focus is on finding adaptive schemes that will be effective on a wide variety of discrete problems. Opportunities abound for graduate and undergraduate students interested in programming, analyzing or conducting experiments.

Project Title: Neural network analysis of images and other data
Faculty Director: J. Daniel Bourland, PhD (bourland@wfubmc.edu) Associate Professor and Head, Physics Section; Department of Radiation Oncology, Wake Forest University School of Medicine; Winston-Salem, North Carolina 27157. Tel: 336-716-2987 Fax: 336-716-7837
Subject Area: Medical applications
Level: Graduate Students

Project Description: We have the need for computer science graduate students for a couple of projects. One that we expect will be funded in August (Whitaker grant proposal by one of my new faculty physicists, Mike Munley, PhD) is for neural network analysis of images and other data for prediction of tissue injury. Similar projects at the MS level, and perhaps the PhD, exist as well.

Project Title: Processor scheduling based on multiple resource needs
Faculty Director: Daniel Cañas
Subject Area: Operating Systems
Level: Graduate Students / Senior level undergraduate.

Project Description: Operating Systems assign the most sought resource, the CPU, mainly, by computing CPU time of a process. Processes require other resources which are generally ignored by the OS scheduler, like memory, I/O, network requests, etc. A metric for computing the resource needs of a process should be part of the scheduler. This project requires the analysis the following:

Project Title: Biological Modeling using Artificial Intelligence Approaches
Faculty Director: Turkett
Subject Area: Computational Biology
Level: Undergraduate and Graduate Students
Project Description: One of the seminal questions in computational systems biology is how to reconstruct the protein signaling pathways that control cellular responses.  A number of biochemical and biological experiments provide mechanisms for measuring genes and proteins within the cell and how those genes and proteins respond to a perturbance.  Given these observations, algorithms can be developed to hypothesize networks that are likely to have generated such data.   This work is interested in exploring those ‘reverse engineering’ algorithms which are motivated by AI techniques, primarily Bayesian network modeling and Support Vector Machine modeling.  There are open questions concerning both the modeling (learning) algorithms themselves, as well as how to accurately pre-pre-process the biological data to prepare for modeling and how to analyze the results of modeling.

Project Title: Improving Run-Time Reasoning Under Uncertainty
Faculty Director: Turkett
Subject Area: Artificial Intelligence
Level: Undergraduate and Graduate Students
Project Description: Reasoning under uncertainty involves decision making when one is faced with both uncertainty in the outcome of actions and one’s own state in the world.  For many robotic and software agent domains, uncertainty is an inescapable component of the domain.   A number of algorithms have been developed to off-line find optimal policies on how to act in these domains, but these algorithms are extremely computationally expensive.  Approximate reactive run-time planning is also difficult because of the extensive look ahead required to develop high quality plans.   This work is to extend and develop new algorithms that examine how additional run-time information, such as reachability, new data-structures for searching, and information sharing between reasoners can be used to reduce the costs of high quality run-time planning under uncertainty.   

For information on the research projects of Reynolds Professor Dr. Robert Plemmons, please see http://www.wfu.edu/~plemmons/.  His research area covers image reconstruction, adaptive optics, and related fields.

For information on the research projects of Reynolds Professor Jacquelyn S. Fetrow, please see http://www.wfu.edu/~fetrowjs/.  Her research areas include protein structure/function relationships; relationship between protein motion, protein function, and drug or inhibitor binding; structure-based drug discovery.