Computer Security
I'm currently working on the applications of soft computing to solve a variety of problems. Of particular interest are those related to computer security (including wireless) and machine learning.
One of my visiting graduate students, Mr. Hongwei Mo, worked on the use of artificial immune systems to document classification.
One of my recent graduate students, Dr. Yihua Liao, worked on the issue of computer security from the point of view of insider threat.
Another student, Ms. Na Tang is applying machine learning methods to solve the so-called inverse problems arising in data poor domains such as the health care field.
Another student, Mr. Khaled Labib is applying statistical methods to study and distinguish traffic patterns during a denial of service attack.
One of my undergraduate interns, Mr. Dustin Puim developed a method of evolving a neural net capable of playing checkers without being told what the rules of the game are.
One of my post-docs, Dr. Wenjie Hu is working on the issue of computer intrusions using a robust support vector machine he developed during his doctoral dissertation.
In a 2000 Ph. D. dissertation, my student Dr. Subhasis Saha developed wavelet-based lossy and lossless compression algorithms
In collaboration with my former student Dr. Kenrick Mock, I have explored the possibility of using genetic algorithms and case based reasoning to develop information filters to filter data coming on the Internet.
In a 1995 Ph. D. dissertation, my student Dr. Walter Cedeno developed the MNC (Multi-Niche Crowding) Genetic Algorithm for finding the peaks of a multi-modal function and applied that technique to assemble the restriction-fragments using the data from chromosome 19 of the Human Genome Project.
In a 1995 dissertation, my student Dr. Ivan Howitt developed a growing algorithm for radial basis function neural networks and applied this method to solve difficult problems arising in the context of digital communications.
In a 1995 Ph. D. dissertation, my student Dr. Jay Smart used simulated annealing to study the difficult problem of visualizing the interdependencies of complex software systems by developing the most desirable "layout" of the dependency graphs.
In a 1995 Ph. D. dissertation, my student Dr. Daniel Styer used reinforcement learning to study the difficult problem of balancing the musculo-skeletal model of a standing human.
In a 1992 Ph. D dissertation, my student Dr. Gyu-Sang Jang solved the difficult problem of discriminating earthquakes from underground nuclear explosions using artificial neural networks.
In a 1981 Ph. D dissertation, my student Dr. Art Raefsky solved a two-point boundary value problem using the finite-element method.