Learning Machines for Clinical Applications
Speaker: Magnus Boman, KTH
Title: Learning Machines for Clinical Applications
A learning machine is an autonomous self-regulating open reasoning machine system that actively learns in an unsupervised (or semi-supervised) and decentralised manner. This talk discusses the case of learning machines for prediction tasks. In Internet-based psychiatry, patients follow a structured programme of cognitive therapy. The learning machine task is here to predict, as early as possible and with as high accuracy as possible, which patients are in need of additional support to successfully complete the programme. In the talk, I will highlight some of the methodological challenges of cross-disciplinary work and of introducing learning machines into a real-life clinical patient flow.
Magnus Boman is a professor of Intelligent Software Services at KTH. He is a part of the Model-Based Computing Systems research group at SCS/ICT, with David Broman and Christian Schulte. With a dozen years in clinical applications of AI, mostly as a computational epidemiologist, he has in recent years turned towards AI and data analytics for psychiatric care, psychological emotion expression, and cognitive decline.