NLM DIR Seminar Schedule

Abstract:

Millions of people daily depend on searching PubMed literature. Providing a reliable, comprehensive, and effective search is crucial for the medical community. To power PubMed literature search countless processes are deployed on the background. In this talk I will highlight several NLP algorithms developed in-house that have been integrated into PubMed search environment. These methods make a difference in how PubMed search handles synonyms to provide more comprehensive results. They define how PubMed recognizes and processes single citation queries, those targeted at finding specific publications. The most recently developed techniques are targeted at monitoring the validity of relevance ranking results obtained through systematic updates of the Relevance Ranking algorithm. Throughout this work, we have contributed and continue to contribute to the research and development of PubMed literature search capabilities.