Fill in the blanks: Rethinking Interpretability in vision

TL;DR

This paper proposes a novel, model-agnostic approach to interpretability in vision models by analyzing how they fill in masked images, revealing consistent input structure patterns and enhancing explainability tools.

Contribution

It introduces a new perspective on vision model interpretability by examining fill-in-the-blank behavior, which is more adaptable across models than traditional saliency-based methods.

Findings

Consistent patterns in model fill-in behavior across datasets

Potential integration as a model-agnostic explainability tool

Improved understanding of learned input structures

Abstract

Model interpretability is a key challenge that has yet to align with the advancements observed in contemporary state-of-the-art deep learning models. In particular, deep learning aided vision tasks require interpretability, in order for their adoption in more specialized domains such as medical imaging. Although the field of explainable AI (XAI) developed methods for interpreting vision models along with early convolutional neural networks, recent XAI research has mainly focused on assigning attributes via saliency maps. As such, these methods are restricted to providing explanations at a sample level, and many explainability methods suffer from low adaptability across a wide range of vision models. In our work, we re-think vision-model explainability from a novel perspective, to probe the general input structure that a model has learnt during its training. To this end, we ask the question: "How would a vision model fill-in a masked-image". Experiments on standard vision datasets and pre-trained models reveal consistent patterns, and could be intergrated as an additional model-agnostic explainability tool in modern machine-learning platforms. The code will be available at \url{https://github.com/BoTZ-TND/FillingTheBlanks.git}