Homer’s Odyssey described mythical sea monsters sited on opposite sides of the narrow strait separating what is now the island of Sicily from the Italian mainland. A rocky shoal (Scylla) and a whirlpool (Charybdis) are real maritime hazards to this day, forcing sailors to balance avoiding one by passing closer to the other.
This metaphor captures how many of us who are not data science experts feel about machine learning in laboratory medicine. We know it’s here; we know it’s important, but how do we bring it into our labs correctly? The plenary session on Wednesday, December 16 at 9:00 am Central, “Between Scylla and Charybdis: Navigating the Complex Waters of Machine Learning in Laboratory Medicine,” deals with these concerns directly.
Ulysses Balis, MD, describes the real-world experience of implementing machine learning-based methods in his clinical laboratory at the University of Michigan. He discusses how laboratorians should not let “fear of the unknown” prevent exploring and leveraging these new technologies.
In fact, Balis points out that “the first principles underlying most data analysis techniques in machine learning are close cousins of the statistical methods we use in clinical labs on a routine basis.” He explains how labs can gain confidence through access to example data sets and by using machine learning solutions that have already been validated.
Balis’ expertise in pathology informatics began with his residency experience at the University of Utah which supports ARUP laboratories. He is a board-certified pathology informaticist and currently serves as the director of the division of pathology informatics, in the department of pathology at the University of Michigan, one of the few such academic information technology divisions operating wholly within a pathology department.
It is Balis’ intent that attendees of this plenary leave the presentation with “the absolute confidence that they can assimilate this new technology and knowledgebase for their effective use within their own lab settings.”