R-Cut: Enhancing Explainability in Vision Transformers with Relationship Weighted Out and Cut
This work addresses the need for better trust and understanding in AI models for users in computer vision, particularly in applications like automatic driving danger alerts, though it appears incremental in improving existing explainability techniques.
The paper tackles the problem of enhancing explainability in Transformer-based image classification models by introducing R-Cut, a method that generates dense class-specific visual explainability maps. The results show significant improvement over previous methods on datasets like ImageNet and LRN, with extensive experiments validating the approach.
Transformer-based models have gained popularity in the field of natural language processing (NLP) and are extensively utilized in computer vision tasks and multi-modal models such as GPT4. This paper presents a novel method to enhance the explainability of Transformer-based image classification models. Our method aims to improve trust in classification results and empower users to gain a deeper understanding of the model for downstream tasks by providing visualizations of class-specific maps. We introduce two modules: the ``Relationship Weighted Out" and the ``Cut" modules. The ``Relationship Weighted Out" module focuses on extracting class-specific information from intermediate layers, enabling us to highlight relevant features. Additionally, the ``Cut" module performs fine-grained feature decomposition, taking into account factors such as position, texture, and color. By integrating these modules, we generate dense class-specific visual explainability maps. We validate our method with extensive qualitative and quantitative experiments on the ImageNet dataset. Furthermore, we conduct a large number of experiments on the LRN dataset, specifically designed for automatic driving danger alerts, to evaluate the explainability of our method in complex backgrounds. The results demonstrate a significant improvement over previous methods. Moreover, we conduct ablation experiments to validate the effectiveness of each module. Through these experiments, we are able to confirm the respective contributions of each module, thus solidifying the overall effectiveness of our proposed approach.