Rethinking uncertainty: A Bayesian perspective on machine learning

While modern deep learning and large-scale AI models deliver unprecedented predictive power, they frequently struggle with uncalibrated confidence, opaque black-box dynamics, and a lack of reliable uncertainty estimation. This workshop examines how a Bayesian perspective offers a principled framework to quantify model uncertainty, incorporate domain priors, and build safer, more trustworthy systems.

You will also be given a demo on how to do a basic project using a Bayesian framework, and then work on your own version of it.

Cover Image: Derived from a digital capture (photo/scan) of the song cover (creator of this digital version is irrelevant as the copyright in all equivalent images is still held by the same party). Copyright held by the record company or the artist. Claimed as fair use regardless., Fair use, https://en.wikipedia.org/w/index.php?curid=556941