SUNY: A Visual Interpretation Framework for Convolutional Neural Networks from a Necessary and Sufficient Perspective

arXiv:2303.00244v2 Announce Type: replace
Abstract: Researchers have proposed various methods for visually interpreting the Convolutional Neural Network (CNN) via saliency maps, which include Class-Activation-Map (CAM) based approaches as a leading family. However, in terms of the internal design logic, existing CAM-based approaches often overlook the causal perspective that answers the core "why" question to help humans understand the explanation. Additionally, current CNN explanations lack the consideration of both necessity and sufficiency, two complementary sides of a desirable explanation. This paper …

abstract arxiv class cnn convolutional convolutional neural network convolutional neural networks cs.ai cs.cv design family framework however interpretation logic map maps network networks neural network neural networks perspective researchers terms type via visual