Template Image
Neural Networks
Offered in Fall 2005, the course started with the concept of classification, learning paradigms - Supervised learning, Unsupervised learning, Reinforced learning but discussed the former two models majorly. Each paradigm explained learning mechanisms broadly classified to four groups such as perceptron learning technique, Hebbian, Backtracking, Bayesian. This was followed by discussion on different clustering techniques useful for learning and classification.
Programming assignments were done in C, and MATLAB. Following are list of mini-projects implemented in class. Each project has a description, source code, screen shots and possible working demo.
Independent Component Analysis
Given a source audio signal formed by mixing two independent audio signals. The goal was to separate two source singals. See here to know more on this Description.
Principal Component Analysis
Given a source audio signal formed by mixing two independent audio signals. The goal was to separate two source singals. See here to know more on this Description.
Finding Classification Parameters Of A Live Web Survey
Given a source audio signal formed by mixing two independent audio signals. The goal was to separate two source singals. See here to know more on this Description.