Distribution Shift as Underspecification, and What We Might Do About It

Distribution Shift as Underspecification, and What We Might Do About It

Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only! Grab it Explore the concept of distribution shift as underspecification in deep learning and potential solutions in this hour-long lecture by Chelsea Finn from Stanford University, presented at the Deep Learning Theory Workshop and Summer School hosted by the Simons Institute. Gain insights into this critical challenge in machine learning and discover strategies for addressing it to improve model performance and generalization.