Update Rules of Pearl and Jeffrey in Probabilistic Updating

In this 56-minute Topos Institute Colloquium talk, explore the contrasting probabilistic updating rules developed by Pearl and Jeffrey. Understand how these approaches differ when updating with multiple data items in learning scenarios. Learn how Jeffrey's rule forms the foundation of the Expectation Maximization algorithm in machine learning. Delivered by Bart Jacobs on April 17, 2025, this presentation provides insights into fundamental probabilistic updating mechanisms and their practical applications in computational learning systems.