Why do These Match? Explaining the Behavior of Image Similarity Models

TL;DR

This paper introduces SANE, a novel method for explaining image similarity models by pairing saliency maps with attributes, enhancing interpretability and attribute recognition across diverse datasets.

Contribution

The paper presents SANE, a new explanation technique tailored for image similarity models that combines saliency maps with attribute explanations, addressing the unique challenges of this task.

Findings

SANE provides explanations that reveal additional information beyond saliency maps.

SANE improves attribute recognition performance.

The method generalizes well across diverse datasets.

Abstract

Explaining a deep learning model can help users understand its behavior and allow researchers to discern its shortcomings. Recent work has primarily focused on explaining models for tasks like image classification or visual question answering. In this paper, we introduce Salient Attributes for Network Explanation (SANE) to explain image similarity models, where a model's output is a score measuring the similarity of two inputs rather than a classification score. In this task, an explanation depends on both of the input images, so standard methods do not apply. Our SANE explanations pairs a saliency map identifying important image regions with an attribute that best explains the match. We find that our explanations provide additional information not typically captured by saliency maps alone, and can also improve performance on the classic task of attribute recognition. Our approach's ability to generalize is demonstrated on two datasets from diverse domains, Polyvore Outfits and Animals with Attributes 2. Code available at: https://github.com/VisionLearningGroup/SANE