Using the Claim Framework to Support Evidence Synthesis and Discovery
Brown Bag Lecture by Dr. Catherine L. Blake | 5/16/2017 11AM – 12PM | 7th Floor Conference Room, Bldg 38A
Abstract: Massive increases in electronically available text have spurred a variety of natural language processing methods to automatically identify relationships from text; however, much of the existing work focuses on either bioinformatics (such as gene–protein relationships) or clinical informatics (such as treatment–disease relationships). This talk introduces the Claim Framework that reflects how authors across biomedical spectrum communicate findings in empirical studies. The Claim Framework captures different levels of evidence by differentiating between explicit and implicit claims, and by capturing results reported as correlations, comparisons, and observations. The talk will introduce the framework and the accompanying automated methods to identify claims automatically. In particular, the automated methods to identify specific facets involved in a comparison claim will be described and a case study from diabetes will be used to illustrate how the facets extracted from comparison claims can be used to populate a tabular summary of evidence. Such a summary supports both evidence synthesis and discovery uncovers by revealing where the literature disagrees and where there are gaps.
The presentation will draw on material in the following papers:
Lucic, A. and Blake, C. (2016), Improving Endpoint Detection to Support Automated Systematic Reviews. American Medical Informatics Association Symposium, Nov 12 – 16, Chicago, IL. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5333237/
Blake, C. and Lucic, A. (2015) An automated approach to identify endpoints to support the systematic review process. Journal of Biomedical Informatics. 56:42-56, Available online August 2015 http://www.sciencedirect.com/science/article/pii/S1532046415000830
Blake, C. (2010) Beyond genes, proteins, and abstracts: Identifying scientific claims from full-text biomedical articles, Journal of Biomedical Informatics, 43(2):173-189. http://www.sciencedirect.com/science/article/pii/S1532046409001476
Bio: Catherine L. Blake, PhD, joined the Lister Hill National Center for Biomedical Communications (LHNCBC) in September 2016. Dr. Blake received her doctorate degree in information and computer science from the University of California, Irvine in 2003. While at NLM, she is on sabbatical leave from the University of Illinois at Urbana Champaign, where she is an associate professor in the School of Information Sciences and associate director of the Center for Informatics Research in Science and Scholarship. At LHNCBC, she is pursuing research that exploits semantic natural language process to automatically synthesize evidence from scientific articles. Her mentor is Thomas Rindflesch, PhD.