Session II Scientific Publishing and Semantic Web Moderator
Session II: Scientific Publishing and Semantic Web Moderator: Alan R. Aronson W 3 C Semantic Web for Life Sciences Workshop October 27, 2004
Foundations of Semantic Text Processing at NLM Alan R. Aronson (National Library of Medicine) Urchin RSS / The Urchin/Kowari Project Ben Lund, David Wood (Nature Publishing Group, Tucana Technologies) Semantic Web and Elsevier Marc Krellenstein (Elsevier) Semantic Web for Data Interpretation & Integration: Lessons Learned from Scientific Publishing and the Distributed Annotation System Steve Chervitz (Affymetrix)
Foundations of Semantic Text Processing at NLM Alan R. Aronson, Ph. D National Library of Medicine W 3 C Semantic Web for Life Sciences Workshop October 27, 2004
Outline • Unified Medical Language System (UMLS) Knowledge Sources • The Meta. Map Program • The NLM Indexing Initiative • Sem. Rep (Semantic Representation)
The Unified Medical Language System • UMLS Knowledge Sources • Metathesaurus • Semantic Network • SPECIALIST Lexicon • Metamorpho. Sys (Metathesaurus subset extraction) • Lexical/spelling tools (lvg, norm, Gspell) • Knowledge Source Server
Meta. Map • Maps text to the Metathesaurus • • • Parse text into phrases Generate word variants Retrieve Metathesaurus candidates Evaluate candidates against text phrases Form final mapping • Linguistically rigorous • Partial matching • Web interface and Java-based application
NLM Indexing Initiative (II) • Investigate automated and semi-automated indexing methodologies • Develop methods that result in acceptable retrieval performance • Concept-based algorithms • Extensive use of UMLS resources • Medical Text Indexer (MTI), a tool for • semi-automated assistance in MEDLINE indexing • automatic indexing of some abstracts collections
Sem. Rep • Family of programs to extract semantic relationships from biomedical text • • • Sem. Rep (the progenitor) Arbiter (binding relationships) EDGAR (drug-gene relationships) Sem. Spec (hypernymic propositions) Sem. Gen (etiology of genetic diseases)
Language and Meaning Language Words Syntactic Structure Predicates Arguments World Model Relations Meaning Entities Semantic Interpretation Semantic Relation(Concept, Concept)
Lexical Look-up and Tagger aggressive combination chemotherapy in adj noun prep the management of hypercalcemic renal failure det noun prep adj noun
Parser mod NP head prep aggressive combination chemotherapy in adj noun prep det NP head prep mod head the management of hypercalcemic renal failure det noun prep adj noun
Meta. Map mod NP head prep aggressive combination chemotherapy in adj noun prep det NP head prep mod head the management of hypercalcemic renal failure det noun prep adj noun Drug Therapy, Combination Kidney Failure topp dsyn Therapeutic or Preventive Procedure Disease or Syndrome
Sem. Rep mod NP head prep aggressive combination chemotherapy in adj noun prep det NP head prep mod head the management of hypercalcemic renal failure det noun prep adj noun Drug Therapy, Combination Kidney Failure topp dsyn Dependency grammar applies syntactic constraints for nominalization
Sem. Rep mod NP head prep aggressive combination chemotherapy in adj noun Drug Therapy, Combination topp prep det NP head prep mod head the management of hypercalcemic renal failure det noun TREATS phsu-TREATS-dsyn prep adj noun Kidney Failure dsyn medd-TREATS-dsyn topp-TREATS-inpo topp-TREATS-sosy topp-TREATS-anab Match semantic types between arguments and Semantic Network
NLM Web Pointers • UMLS Knowledge Source Server: http: //umlsks. nlm. nih. gov/ • Semantic Knowledge Representation Project: http: //skr. nlm. nih. gov/ • NLM Indexing Initiative: http: //ii. nlm. nih. gov/
- Slides: 15