Our #85 Webinar Meetup will be hosted by Inês Pereira and dedicated to the topic of “Model interpretability in healthcare”. This Meetup will be held exclusively online.
The schedule for the webinar is the following:
• 18:30 – 18:45: Opening the meeting
• 18:45 – 19:30: Model interpretability in healthcare by Inês Pereira
• 19:30 – 19:45: Q&A
• 19:45: Closing
See you there!
This webinar is sponsored by Bosch (https://www.bosch.pt/)!
Model interpretability in healthcare
In recent years, hype and, in particular, claims of superhuman performance have pervaded the literature on machine learning models applied to healthcare. This has led to important debates on the topics of model development and validation, dealing with bias, and addressing the issue of model interpretability, with heated debates taking place both live and online. When it comes to model interpretability, there still is no mathematical agreed-upon definition. Non-mathematical definitions have however been proposed, as well as many techniques to peek inside a model’s inner workings. In this talk, we will review the concepts of interpretability and explainability and discuss the need for interpretability with concrete examples. In addition, under the notion that interpretability can also serve as a means to more granular diagnoses, prominent modelling approaches used in the field of Computational Psychiatry will also be presented.
Inês Pereira holds both a Medical degree from the NOVA Medical School of Lisbon and an MSc in Neural Systems and Computation from UZH and ETH Zurich, Switzerland.
Having worked as a Neurology resident in Berlin, Germany, she is now pursuing a PhD at the Translational Neuromodeling Unit, in Zurich, Switzerland, focusing on generative modelling of brain connectivity as a tool for differential diagnosis and prediction in Computational Psychiatry and Computational Neurology.
Other characteristics: avid language learner and part-time blogger. Personal webpage: https://inespereira.com/