contacttdeoskar AT inf DOT ed DOT ac DOT uk
Tel: +44 (0) 131 651 5643
I have moved to the University of Amsterdam, as of Feb 2016. Please click here to find my new homepage.
my research in a nutshellI am interested in statistical models of language, particularly of syntax. I have worked on building statistical parsing models whose representations are more linguistically sophisticated that typical parsers. Due to the data-sparseness associated with supervised models with complex representations, I am specifically interested in harnessing unlabeled data for these by using semi-supervised learning techniques.
In the GramPlus project, we are engaged in solving challenges related to the generalization of statistical CCG grammars in order to make them both broad-coverage and accurate.
Additionally, I am interested in the syntax and semantics of languages spoken in India, specifically Hindi and Marathi.
recent and upcoming activities
linksCombinatory Categorial Grammar
Reading Group Probabilitic Models of Language
Reading Group Machine Learning for NLP