Natural Language Processing, Machine Translation and Artificial Intelligence
Domain of Expertise
Standardization, Research Areas, Linguistic Resources & Tools, Application Showcase
He has worked on Computational Paninian Grammar (CPG) framework which is linguistically elegant and computationally efficient. It is a dependency grammar based framework, particularly well suited for morphologically rich and free word-order languages. Constraint-based parser for it uses integer programming and bipartite graph matching. He has also worked on machine learning of parsing, reaching high accuracy even with a small treebank. The CPG framework now forms the basis of work on Indian language processing, including machine translation among Indian languages.