Title: Producing Joint Decisions and Explanations with CNNs
Abstract:
CNNs have shown remarkable performance on tasks like classification and detection. However, with the growing demand of the right to explanation, especially in highly-regulated areas like healthcare and criminal justice, generating accurate predictions is often not enough. These models have to be explainable, i.e., understandable to humans, which entails being able to present the reasons behind their decisions. While most of the literature addresses post-model methods, this talk will focus on an in-model CNN architecture, composed by an explainer and a classifier. The model is trained end-to-end, with the classifier taking as input not only images from the dataset but also the explainer’s resulting explanation, thus allowing for the classifier to focus on the relevant areas of such explanation.
