Teaching Humans Subtle Differences with DIFFusion

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chiquier, Mia, Avrech, Orr, Gandelsman, Yossi, Feng, Berthy, Bouman, Katherine, Vondrick, Carl
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908365158875136
author Chiquier, Mia
Avrech, Orr
Gandelsman, Yossi
Feng, Berthy
Bouman, Katherine
Vondrick, Carl
author_facet Chiquier, Mia
Avrech, Orr
Gandelsman, Yossi
Feng, Berthy
Bouman, Katherine
Vondrick, Carl
contents Scientific expertise often requires recognizing subtle visual differences that remain challenging to articulate even for domain experts. We present a system that leverages generative models to automatically discover and visualize minimal discriminative features between categories while preserving instance identity. Our method generates counterfactual visualizations with subtle, targeted transformations between classes, performing well even in domains where data is sparse, examples are unpaired, and category boundaries resist verbal description. Experiments across six domains, including black hole simulations, butterfly taxonomy, and medical imaging, demonstrate accurate transitions with limited training data, highlighting both established discriminative features and novel subtle distinctions that measurably improved category differentiation. User studies confirm our generated counterfactuals significantly outperform traditional approaches in teaching humans to correctly differentiate between fine-grained classes, showing the potential of generative models to advance visual learning and scientific research.
format Preprint
id arxiv_https___arxiv_org_abs_2504_08046
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Teaching Humans Subtle Differences with DIFFusion
Chiquier, Mia
Avrech, Orr
Gandelsman, Yossi
Feng, Berthy
Bouman, Katherine
Vondrick, Carl
Computer Vision and Pattern Recognition
Scientific expertise often requires recognizing subtle visual differences that remain challenging to articulate even for domain experts. We present a system that leverages generative models to automatically discover and visualize minimal discriminative features between categories while preserving instance identity. Our method generates counterfactual visualizations with subtle, targeted transformations between classes, performing well even in domains where data is sparse, examples are unpaired, and category boundaries resist verbal description. Experiments across six domains, including black hole simulations, butterfly taxonomy, and medical imaging, demonstrate accurate transitions with limited training data, highlighting both established discriminative features and novel subtle distinctions that measurably improved category differentiation. User studies confirm our generated counterfactuals significantly outperform traditional approaches in teaching humans to correctly differentiate between fine-grained classes, showing the potential of generative models to advance visual learning and scientific research.
title Teaching Humans Subtle Differences with DIFFusion
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2504.08046