MAVias: Mitigate any Visual Bias

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Sarridis, Ioannis, Koutlis, Christos, Papadopoulos, Symeon, Diou, Christos
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915624605712384
author Sarridis, Ioannis
Koutlis, Christos
Papadopoulos, Symeon
Diou, Christos
author_facet Sarridis, Ioannis
Koutlis, Christos
Papadopoulos, Symeon
Diou, Christos
contents Mitigating biases in computer vision models is an essential step towards the trustworthiness of artificial intelligence models. Existing bias mitigation methods focus on a small set of predefined biases, limiting their applicability in visual datasets where multiple, possibly unknown biases exist. To address this limitation, we introduce MAVias, an open-set bias mitigation approach leveraging foundation models to discover spurious associations between visual attributes and target classes. MAVias first captures a wide variety of visual features in natural language via a foundation image tagging model, and then leverages a large language model to select those visual features defining the target class, resulting in a set of language-coded potential visual biases. We then translate this set of potential biases into vision-language embeddings and introduce an in-processing bias mitigation approach to prevent the model from encoding information related to them. Our experiments on diverse datasets, including CelebA, Waterbirds, ImageNet, and UrbanCars, show that MAVias effectively detects and mitigates a wide range of biases in visual recognition tasks outperforming current state-of-the-art.
format Preprint
id arxiv_https___arxiv_org_abs_2412_06632
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MAVias: Mitigate any Visual Bias
Sarridis, Ioannis
Koutlis, Christos
Papadopoulos, Symeon
Diou, Christos
Computer Vision and Pattern Recognition
Mitigating biases in computer vision models is an essential step towards the trustworthiness of artificial intelligence models. Existing bias mitigation methods focus on a small set of predefined biases, limiting their applicability in visual datasets where multiple, possibly unknown biases exist. To address this limitation, we introduce MAVias, an open-set bias mitigation approach leveraging foundation models to discover spurious associations between visual attributes and target classes. MAVias first captures a wide variety of visual features in natural language via a foundation image tagging model, and then leverages a large language model to select those visual features defining the target class, resulting in a set of language-coded potential visual biases. We then translate this set of potential biases into vision-language embeddings and introduce an in-processing bias mitigation approach to prevent the model from encoding information related to them. Our experiments on diverse datasets, including CelebA, Waterbirds, ImageNet, and UrbanCars, show that MAVias effectively detects and mitigates a wide range of biases in visual recognition tasks outperforming current state-of-the-art.
title MAVias: Mitigate any Visual Bias
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.06632