Title: Machine Learning for Extraction and Retrieval of Biomedical Text DataAbstract: An extensive amount of biomedical information is available only as natural language text, which inhibits our ability to find and analyze it properly. The underlying reason for the preponderance text data is directly tied to its computational difficulty: natural language is a quick and expressive means of encoding ideas. Simply put, natural language is awesome, but also terrible. Strategies for how to best leverage natural language data, then, must include a wide range of sophisticated techniques. This talk will walk through a course of methods for biomedical text, from low-level language understanding of clinical sentences, to combining information from dozens of patient notes, to supporting clinical decisions through search engines. Throughout, emphasis will be placed on the advanced machine learning models dominating these tasks.Bio: Kirk Roberts, PhD, is an assistant professor at the UTHealth School of Biomedical Informatics. He is an expert on biomedical natural language processing (NLP) and information retrieval (IR). He is particularly interested in developing robust methods that application-specific annotation and/or system development. He is an organizer of the Clinical NLP workshop series, numerous TREC biomedical IR tracks, and is co-leader of the JAMIA student editorial board.