Research

At an higher level, my Research interest is :

• Probabilistic perspective of Machine Learning and Signal Processing. In particular, my research encompass Neural Networks, faster second order training algorithms, high dimensional feature selection and Boltzmann Machines.

My Research goal is :

• Develop and test classifiers based on new Kernel Machine methods that involved novel embedded feature selection and feature generation.

This is what keeps me buys in a day ( or night !!! ). Besides this in future, i am planning to foray into the application field of Machine Learning so that after we have a decent algorithm, I want to make it as a useful and wonderful demo. not just usual charts and curves. Probably i am too much impressed by Machine Learning that Matters and Impact of Machine Learning in real world.

Building Better Support Vector Machines

Deep Learning

Second order Training algorithm for Kernel Machine

Development of L-1 based subspace methods and Viola Jones face detection algorithm

Designing a advance Multilayer Perceptron (MLP) algorithm for Approximation and Classification problems