PhD Dissertation: Developing and Applying an Integrated Semantic Framework for Natural Language Understanding

There is much active research in Computational Linguistics on deep semantics of natural language.
Various tools, methods and technologies have been developed for extracting and manipulating information from textual data.
Although these technologies have proven to be useful for many purposes, they are mainly based on separated modular architectures and thus are subjected to improvement.

In this research, I will study the state-of-the-art approaches for computational semantic analysis and introduce a new integrated semantic representation for such purposes.
By integrating information from various perspectives such as lexical semantics, structural semantics, and knowledge bases, we expect to achieve better analyses in core tasks such as word-sense disambiguation, parse ranking  and knowledge extraction.
A prototype system will be developed usingHead-driven Phrase Structure Grammar (HPSG: Sag, Wasow, & Bender, 2003) Minimal Recursion Semantics (MRS: Copestake, Flickinger, Sag, & Pollard, 2005) and WordNet (WN: Fellbaum, 1998).