Abstract: Knowledge-Based Semantic Interpretation, By Dr. Thomas Rindflesch
|
|||||
|
|
|
||||
|
||||||||||||
|
Abstract posted for the May 20-21 AGI Workshop: Knowledge-Based Semantic Interpretation for Summarizing Biomedical Text Dr. Thomas Rindflesch Marcelo Fiszman Halil Kilicoglu François Lang National Library of Medicine  PPT: Knowledge-Based Semantic Interpretation for Summarizing Biomedical Text SemRep is a natural language processing system being developed to recover semantic predications from biomedical text using underspecified syntactic analysis and structured domain knowledge. The SPECIALIST Lexicon and a stochastic tagger support the identification of simple noun phrases and verb groups, while domain knowledge is provided by components of the Unified Medical Language System (UMLS). During interpretation, simple noun phrases functioning as referring expressions are mapped to concepts in the Metathesaurus, and syntactic phenomena that “indicate” semantic relations are mapped to predicates in the Semantic Network. Syntactic constraints on argument identification are controlled by an underspecified dependency grammar. SemRep serves as the basis for several ongoing research initiatives in biomedical information management, including automatic summarization in the semantic abstraction paradigm. Predications generated by SemRep from a set of Medline citations on a specified disorder topic are summarized into a conceptual condensate of the input text. The method relies on four principles, which transform a list of predications into a summarized version: All predications are subject to frequency of occurrence constraints (saliency principle), while relevance and connectivity determine that both core predications on the topic and additional “useful” predications should be included. Finally, novelty excludes predications that the user likely already knows. It is increasingly challenging for researchers, health professionals, and consumers to effectively exploit the extensive textual resources provided by the National Library of Medicine. We are developing a Web application that exploits automatic summarization in cooperation with PubMed to pinpoint the most relevant information in large document sets. The summary is displayed in graphical format with links to the text that generated it as well as connections to supporting structured knowledge such as the UMLS Metathesaurus, Entrez Gene, and the Genetics Home Reference. Important Links: Main Workshop Website: http://www.agiri.org/workshop Directions/Hotel: http://www.agiri.org/directions.htm Workshop Schedule: http://www.agiri.org/schedule.htm Printable Version / Handout: http://www.agiri.org/workshop/AGIRI_Workshop_2006.pdf |
    |
|
AGIRI · Forums · Discussion Lists · Wiki |  |
AGIRI - Artificial General Intelligence Research Institute Copyright © 2001-2007 :: User Agreement :: Discussion Lists :: Contact |