Towards Higher-Order and Disentangled Explainable AI

Explore advanced concepts in Explainable AI through this conference talk by Gregoire Montavon from Freie Universität Berlin. Delve into the evolution of explanation techniques, moving beyond first-order methods like LRP to more sophisticated approaches. Discover how to identify joint contributions of feature collections and decompose explanations into disentangled components. Gain insights into higher-order and disentangled explanations that enhance the expressiveness of AI interpretability methods. Understand the latest developments in making complex nonlinear machine learning models more transparent and interpretable.

Gregoire Montavon - Towards Higher-Order and Disentangled Explainable AI - IPAM at UCLA