Research
My primary area of interest is in Natural Language Processing, although I am interested in general Artificial Intelligence and Machine Learning, Game Playing, etc. as well. I am currently enrolled as a Masters’ student in the Department of Computer Sciences, and aspire to get a PhD degree afterwards.
In my recent Computational Linguistics class (offered by Dr. Katrin Erk), I tried to use dependency based semantic spaces published very recently by Sebastian Pado and Mirella Lapata (Computational Linguistics, 33:2, 2007) for word sense disambiguation. I used this dependency based word vector space instead of a simple co-occurrence based vector space and compared the two systems in word sense disambiguation using a similar approach as Hinrich Schutze in his paper, Automatic Word Sense Discrimination, (Computational Linguistics, 24:1, 1998). A copy of the class project report I submitted is here.
I have also worked earlier in my Neural Networks class with Dr. Risto Miikkulainen. In this class, I tried to extend Q-learning based routing algorithms by trying to implement Probabilistic Confidence-based Dual Reinforcement Q - learning, or PCDRQ routing in short. This was my final project report and I am still interested in taking the project further to see if the new protocol can be implemented and scaled to real-time network routing.