TAME: Attention Mechanism Based Feature Fusion for Generating Explanation Maps of Convolutional Neural Networks
This addresses the problem of interpretability for users of CNNs in applications requiring explainability, though it appears incremental as it builds on existing attention-based explanation techniques.
The paper tackles the lack of explainability in neural networks by introducing TAME, a method that uses a multi-branch hierarchical attention mechanism to generate explanation maps from feature maps in CNNs, showing improvements over previous top-performing methods on models like VGG-16 and ResNet-50 trained on ImageNet.
The apparent ``black box'' nature of neural networks is a barrier to adoption in applications where explainability is essential. This paper presents TAME (Trainable Attention Mechanism for Explanations), a method for generating explanation maps with a multi-branch hierarchical attention mechanism. TAME combines a target model's feature maps from multiple layers using an attention mechanism, transforming them into an explanation map. TAME can easily be applied to any convolutional neural network (CNN) by streamlining the optimization of the attention mechanism's training method and the selection of target model's feature maps. After training, explanation maps can be computed in a single forward pass. We apply TAME to two widely used models, i.e. VGG-16 and ResNet-50, trained on ImageNet and show improvements over previous top-performing methods. We also provide a comprehensive ablation study comparing the performance of different variations of TAME's architecture. TAME source code is made publicly available at https://github.com/bmezaris/TAME