Keywords: Backgrounds, Model Biases, Robustness, Computer Vision
Abstract: We assess the tendency of state-of-the-art object recognition models to depend on signals from image backgrounds. We create a toolkit for disentangling foreground and background signal on ImageNet images, and find that (a) models can achieve non-trivial accuracy by relying on the background alone, (b) models often misclassify images even in the presence of correctly classified foregrounds--up to 88% of the time with adversarially chosen backgrounds, and (c) more accurate models tend to depend on backgrounds less. Our analysis of backgrounds brings us closer to understanding which correlations machine learning models use, and how they determine models' out of distribution performance.
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics
One-sentence Summary: We develop and use a toolkit to investigate models’ use of (and reliance on) image backgrounds.
Supplementary Material: zip
Code: [![github](/images/github_icon.svg) MadryLab/backgrounds_challenge](https://github.com/MadryLab/backgrounds_challenge)
Community Implementations: [![CatalyzeX](/images/catalyzex_icon.svg) 1 code implementation](https://www.catalyzex.com/paper/arxiv:2006.09994/code)
17 Replies
Loading