Natural Language Processing For Intelligent Information Retrieval

Abstract

An ongoing research effort at NLM's Lister Hill National Center for Biomedical Communications concerns natural language processing (NLP) for improved repmentation and retrieval of biomedical information. The primary focus of this research program is the development of SPECIALIST, an experimental system for parsing and accessing biomedical text. The system uses lexical, morphologic, syntactic and semantic information as it builds structured representations of sentences and phrases in biomedical texts. It includes a database retrieval facility to allow for testing and evaluation of natural language processing techniques in information retrieval. The system is currently being augmented with information from the two recently released Unified Medical Language System (Urnm) knowledge sources, Meta-lm and the UMLS Semantic Network. The SPECIALIST system is implemented in Quintus Prolog and C and runs on Sun 4 workstations.