iSchool faculty will present their research at the AMIA (American Medical Informatics Association) Informatics Summit, which will be held from March 21-24 in Chicago. The symposium showcases the latest innovations from the community of biomedical informatics researchers and practitioners.
Associate Professor Halil Kilicoglu will present "Investigating the Impact of Weakly Supervised Data on Text Mining Models of Publication Transparency: A Case Study on Randomized Controlled Trials." According to Kilicoglu, the lack of large quantities of labeled data is a major barrier in developing effective text mining models of biomedical literature. To combat this, his team (PhD students Linh Hoang and Lan Jiang) used weak supervision and data augmentation techniques (Snorkel, UMLS-EDA) to create additional labeled data to improve the models that assess transparency of clinical trial publications.
Affiliate Professor Neil Smalheiser will present the paper, "Testing a Filtering Strategy for Systematic Reviews: Evaluating Work Savings and Recall," which he coauthored with Assistant Professor Jodi Schneider; PhD student Tzu-Kun (Esther) Hsiao; Randi Proescholdt (MS/LIS '20), Menlo College; and Aaron Cohen and Marian McDonagh of Oregon Health and Science University. In the paper, the researchers discuss how they used two machine learning models to filter publication types, in order to make systematic reviews on drug effectiveness more efficient. They found that automated publication type filtering can "potentially provide substantial work savings with minimal loss of included articles." This work was part of the National Institutes of Health (NIH) grant, "Text Mining Pipeline to Accelerate Systematic Reviews in Evidence-Based Medicine," of which Schneider and Smalheiser served as investigators.