Faithful Attention Explainer: Verbalizing Decisions Based on Discriminative Features

Yao Rong, David Scheerer, Enkelejda Kasneci

Research output: Contribution to journalConference articlepeer-review

Abstract

In recent years, model explanation methods have been designed to interpret model decisions faithfully and intuitively so that users can easily understand them. In this paper, we propose a framework, Faithful Attention Explainer (FAE), capable of generating faithful textual explanations regarding the attended-to features. Towards this goal, we deploy an attention module that takes the visual feature maps from the classifier for sentence generation. Furthermore, our method successfully learns the association between features and words, which allows a novel attention enforcement module for attention explanation. Our model achieves promising performance in caption quality metrics and a faithful decision-relevance metric on two datasets (CUB and ACT-X). In addition, we show that FAE can interpret gaze-based human attention, as human gaze indicates the discriminative features that humans use for decision-making, demonstrating the potential of deploying human gaze for advanced human-AI interaction.

Original languageEnglish
Pages (from-to)33-40
Number of pages8
JournalCEUR Workshop Proceedings
Volume3793
StatePublished - 2024
EventJoint of the 2nd World Conference on eXplainable Artificial Intelligence Late-Breaking Work, Demos and Doctoral Consortium, xAI-2024:LB/D/DC - Valletta, Malta
Duration: 17 Jul 202419 Jul 2024

Keywords

  • Explainable AI (XAI)
  • Faithfulness
  • Saliency Map
  • Textual Explanations
  • Visual Explanation

Fingerprint

Dive into the research topics of 'Faithful Attention Explainer: Verbalizing Decisions Based on Discriminative Features'. Together they form a unique fingerprint.

Cite this