A tool for deep learning model interpretability

Abstract: The deep learning (DL) model decision process is famously opaque. However, DL models are not “black boxes” with modern DL interpretability tools, we can obtain useful explanations for DL models. In this talk, we highlight a high-level interpretability approach: intermediate layer probing. We motivate the approach with a couple of case studies from the literature, focusing on concept activation vectors.

TAG-DS is a hybrid seminar and will be available in-person at the UW Mathematics Department as well as online on Zoom. You can find the link to the Zoom meeting here. If you would like to be added to our mailing list, you can do so by visiting this page.