Data Clustering With Spiking Neurons
MARINA KOGAN'S FINAL REPORT
Student research team: Marina Kogan, junior student in 2003/2004, CSI/CUNY
Faculty Mentor: Dr. Natacha Gueorguieva, Associate Professor, Department of CS, CSI/CUNY
School: College of Staten Island/City University of New York (CSI/CUNY)
Address: 2800 Victory Blvd, Staten Island, NY 10314
Data Clustering With Spiking Neurons
Student research team: Marina Kogan, junior student in 2003/2004, CSI/CUNY
Faculty Mentor: Dr. Natacha Gueorguieva, Associate Professor, Department of CS, CSI/CUNY
School: College of Staten Island/City University of New York (CSI/CUNY)
Address: 2800 Victory Blvd, Staten Island, NY 10314
The goal of the project was to become familiar with Spiking Neurons, their biological functions and simulative capabilities. Moreover, constructing viable simulations utilizing the Neural Networks was a projected result as well. Multiple experiments with extensive analysis of the outcomes were expected to be a part of this research project. We can characterize all such stages of research as "reading up", familiarizing, utilizing, synthesizing and analyzing. All such common aspects inherent in real life research were present in our research project. We read and analyzed various academic papers and works, mastered several available Neural Network software applications and libraries, constructed multiple simulations using data structures and utilities of one such software. We constructed six different Neural Network architectures with 3-4 layers of Spiking neurons, differing one from another by the presence / absence of various connections, which could also be bidirectional and plain. The biological parameters, such as weights of synapses and time delays, remained the same from simulation to simulation. Consequently, all the variations in the resulting membrane potential graphs could be attributed to change from one particular architecture to another. We analyzed those graphs and arrived at an enlightening conclusion: our simulation responds in accordance with the expected values of that of a real brain, and therefore, could be used for solving various brain related problems.
We also gained confidence in our analytical power and capability for critical thinking. We exercised and improved our programming skills, which was a great source of confidence as well. As female Computer Science students, this boost of confidence is extremely important as a source of inspiration for the future accomplishments, such as continuing our education into Graduate school, contributing to real-life research, or finding creative and fulfilling employment.
Publications
Project web site:
http://www.cs.csi.cuny.edu/~natacha/Projects/ProjectGrant.htm
http://www.cs.csi.cuny.edu/~natacha/Projects/neurons/