Solving y=f(x) in healthcare with AI

Biography: Dr. Yuan Luo is an Associate Professor at Feinberg School of Medicine in Northwestern University. He is Chief AI Officer at Northwestern University Clinical and Translational Sciences Institute and Institute for Augmented Intelligence in Medicine. He earned his PhD degree from MIT EECS with a math minor and a certificate in Graduate Education of Medical Science (GEMS). His research interests include machine learning, natural language processing, time series analysis, computational phenotyping and integrative genomics, with a focus on biomedical applications. He won the American Medical Informatics Association (AMIA) New Investigator Award in 2020. His PhD Thesis was awarded the inaugural Doctoral Dissertation Award Honorable Mention by AMIA in 2017. He is currently an editor with JAMIA Open, JBI, Plos One, JHIR. He served on AMIA Membership and Outreach Committee. His publications appear in leading journals including Nature Medicine, AJRCCM, JAMIA, JBI, BiB etc. He has published in and served as PC members for top AI and informatics conferences including AAAI, IJCAI, AMIA etc.

Research Overview: The application of artificial intelligence (AI) to the analysis of electronic healthcare data has received extensive attention recently. Multi-modal healthcare big data brings new opportunities and challenges to the application of artificial intelligence. We aim to explore the different modalities of healthcare big data (for example, unstructured clinical records, structured electronic medical case data, medical imaging data, multi-omics data, etc.), and show how to separately and/or jointly mine these data modalities to derive useful information and support clinical decision-making. Our recent work in the development of related AI algorithms has seen applications in data modalities such as textual clinical records, multi-omics, and time series of physiologic variables and laboratory test results. The common theme of these studies is to construct clinical models that can improve the accuracy and interpretability of predictions by exploring and combining relevant information in different data modalities.