Anagha Kulkarni (Joshi) - Research Interests

My research work in Natural Language Processing (NLP) is based on unsupervised approaches.
We are working towards employing unsupervised Word Sense Disambiguation(WSD) techniques to the problems like:
1. Proper Name Discrimination
2. Email Categorization
3. Synonym finding / word clustering

We use and extend SenseClusters for the above problems, which is originally developed by Dr. Ted Pedersen and Amruta Purandare.

Our Scripts which help us create data for our Experiments -
1. 'Headless' and Senseval2 formatted data creator
2. Name Conflate

My Thesis Proposal and the Reading List


I worked as an intern at the Biomedical Informatics (BMI) Division of Mayo Clinic, Rochester over the Summer of 2005. The focus of this internship was on "Cluster Stopping". A copy of my final presentation at Mayo can be accessed here.

Back to HomePage Last Updated: 08 Oct 2005