Learning Visual Explanations for DCNN-Based Image Classifiers Using an Attention Mechanism
This work addresses the need for interpretability in AI for researchers and practitioners using DCNN-based image classifiers, though it is incremental as it builds on existing CAM and attention mechanisms.
The paper tackles the problem of explaining deep convolutional neural network (DCNN) image classifiers by proposing two new learning-based XAI methods, L-CAM-Fm and L-CAM-Img, which use an attention mechanism to generate class activation maps (CAMs) from feature maps or input images, achieving competitive results on ImageNet with a single forward pass at inference.
In this paper two new learning-based eXplainable AI (XAI) methods for deep convolutional neural network (DCNN) image classifiers, called L-CAM-Fm and L-CAM-Img, are proposed. Both methods use an attention mechanism that is inserted in the original (frozen) DCNN and is trained to derive class activation maps (CAMs) from the last convolutional layer's feature maps. During training, CAMs are applied to the feature maps (L-CAM-Fm) or the input image (L-CAM-Img) forcing the attention mechanism to learn the image regions explaining the DCNN's outcome. Experimental evaluation on ImageNet shows that the proposed methods achieve competitive results while requiring a single forward pass at the inference stage. Moreover, based on the derived explanations a comprehensive qualitative analysis is performed providing valuable insight for understanding the reasons behind classification errors, including possible dataset biases affecting the trained classifier.