10/07/2014: I gave a talk on Dirichlet Process and related gaussian mixture model and also presented a faster inference technique using Cholesky decomposition of the covariance matrix (pdf). Update: I also wrote a more detailed tutorial.
My research interests are in the areas of Natural Language Processing, Machine Learning and analysis of social media. The potential of data mining that social media data gives us is fascinating and I am very interested in developing practical applications that would be of benefit to mankind. Here's a project that I have been working on for a while.
I am also interested in usable privacy and have been working on a very interesting project in which we try to understand privacy policies of websites by applying NLP techniques.
Studying the social dynamics of a urban city: I have been associated with the Livehoods project where we try to study the dynamic picture of a city defined by the people who make it a part of their daily life. Presently, I am working on a non parametric model to automagically divide the city into 'neighborhoods' (we are submitting this work to KDD 2014). Please feel free to use my Java implementation of the distance dependent chinese restaurant process. I am presently working on polishing the code (read documenting). But if you are in urgent need, feel freee to email me for help.
Algorithms for NLP
Language and Statistics
Probablistic Graphical Models
A subset of upcoming NLP | IR | ML | AI conferences