Integrating deep and shallow natural language processing components [Elektronische Ressource] : representations and hybrid architectures / eingereicht von Ulrich Schäfer

Integrating Deep and Shallow Natural LanguageProcessing Components – Representations andHybrid ArchitecturesDissertation zur Erlangung des Grades desDoktors der Ingenieurwissenschaften derNaturwissenschaftlich-Technischen Fakulta¨tender Universita¨t des SaarlandesEingereicht von Dipl.-Inform. Ulrich Scha¨ferSaarbru¨cken, 10. Dezember 2006Datum des Promotionskolloquiums: 29. Juni 2007Dekan der Naturwissenschaftlich-Technischen Fakulta¨t I (Mathematik undInformatik): Prof. Dr.-Ing. Thorsten HerfetVorsitzender: Prof. Dr. Andreas ZellerBerichterstattende: Prof. Dr. Hans Uszkoreit, Prof. Dr. Wolfgang WahlsterAkad. Mitarbeiter: Dr. Stephan Busemann2AbstractWe describe basic concepts and software architectures for the integration ofshallow and deep (linguistics-based, semantics-oriented) natural language process-ing (NLP) components. The main goal of this novel, hybrid integration paradigm isimproving robustness of deep processing. After an introduction to constraint-basednatural language parsing, we give an overview of typical shallow processing tasks.We introduce XML standoff markup as an additional abstraction layer that easesintegration of NLP components, and propose the use of XSLT as a standardizedand efficient transformation language for online NLP integration.In the main part of the thesis, we describe our contributions to three hybrid ar-chitecture frameworks that make use of these fundamentals.