You are here
Semantic Knowledge Representation Project: Advanced Information Management for Biomedical Research
It is increasingly challenging for researchers and health professionals to exploit the extensive textual resources provided by NLM. High throughput natural language processing applications can supplement Library information retrieval services such as PubMed. The Semantic Knowledge Representation (SKR) project conducts basic research in symbolic natural language processing based on the UMLS knowledge sources. In addition, project staff develop methodologies and applications for advanced biomedical information management.
A core resource is the SemRep program, which extracts semantic propositions from biomedical text. A facility is under development for extending SemRep to process proposition-modifying information, including speculations, opinions, evidence, and attitudes. Processing is knowledge intensive and relies on the UMLS Metathesaurus and Semantic Network as an ontology underpinning the identification of propositions in biomedical text. SemRep was originally developed for biomedical research. A general methodology is being devised for extending its domain, currently to influenza epidemic preparedness, health promotion, and medical informatics language processing.
SKR efforts support innovative information management approaches in biomedicine, as well as basic research. The Semantic MEDLINE Web application integrates information retrieval, advanced natural language processing, automatic summarization, and visualization into a single Web portal. The project team is using semantic predications to find publications that support critical questions used during the creation of clinical practice guidelines (with support from NHLBI). Other work exploits predications and graph theory for automatic summarization of biomedical text. Significant research is being devoted to developing and applying the literature-based discovery paradigm using semantic predications.