Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

meaningful perturbation

Meaningful perturbation is an interpretability technique in machine learning used to explain decisions made by complex models by applying controlled, visually or semantically coherent modifications to an input and observing the resulting change in the model output. Rather than inspecting the internal parameters or gradients of a network, this approach systematically alters specific regions or features—such as blurring, occluding, or masking portions of an image—to determine which areas cause the largest drop in prediction confidence. By finding the smallest or most influential modification required to change the decision while avoiding artificial artifacts, meaningful perturbation generates human-interpretable attribution maps that pinpoint the exact parts of an input most responsible for the model classification.

1 item

Interpretable Explanations of Black Boxes by Meaningful Perturbation

Interpretable Explanations of Black Boxes by Meaningful Perturbation

Ruth Fong, Andrea Vedaldi

OrganizationsUniversity of Oxford

Why you should read this

Develops a model-agnostic framework that produces faithful visual explanations for black-box image classifiers by optimizing meaningful and testable input perturbations.

As machine learning algorithms are increasingly applied to high impact yet high risk tasks, such as medical diagnosis or autonomous driving, it is critical that researchers can explain how such algorithms arrived at their predictions. In recent years, a number of image saliency methods have been developed to summarize where highly complex neural networks "look" in an image for evidence for their predictions. However, these techniques are limited by their heuristic nature and architectural constraints. In this paper, we make two main contributions: First, we propose a general framework for learning different kinds of explanations for any black box algorithm. Second, we specialise the framework to find the part of an image most responsible for a classifier decision. Unlike previous works, our method is model-agnostic and testable because it is grounded in explicit and interpretable image perturbations.

Added

2026-09-24