paper-with-me

홈 › Papers

Discover the Unknown Biased Attribute of an Image Classifier

2021-04-29 · ICCV 2021 10 · Zhiheng Li, Chenliang Xu

Recent works find that AI algorithms learn biases from data. Therefore, it is urgent and vital to identify biases in AI algorithms. However, the previous bias identification pipeline overly relies on human experts to conjecture potential biases (e.g., gender), which may neglect other underlying biases not realized by humans. To help human experts better find the AI algorithms' biases, we study a new problem in this work -- for a classifier that predicts a target attribute of the input image, discover its unknown biased attribute. To solve this challenging problem, we use a hyperplane in the generative model's latent space to represent an image attribute; thus, the original problem is transformed to optimizing the hyperplane's normal vector and offset. We propose a novel total-variation loss within this framework as the objective function and a new orthogonalization penalty as a constraint. The latter prevents trivial solutions in which the discovered biased attribute is identical with the target or one of the known-biased attributes. Extensive experiments on both disentanglement datasets and real-world datasets show that our method can discover biased attributes and achieve better disentanglement w.r.t. target attributes. Furthermore, the qualitative results show that our method can discover unnoticeable biased attributes for various object and scene classifiers, proving our method's generalizability for detecting biased attributes in diverse domains of images. The code is available at https://git.io/J3kMh.

📄 PDF Abstract BibTeX arXiv:2104.14556

Code (1)

zhihengli-UR/discover_unknown_biases 공식 구현 pytorch

Tasks

AttributeDisentanglement

Similar Papers 제목 키워드 기반

BiasEdit: A Training-Free Bias-Detect-and-Edit Framework for Learning Fair Visual Classifiers

2026-05-27 · Jungwook Seo, Yoonsik Park, Changmin Lee, Sungyong Baik arxiv

Visual data from the Web power image classifiers, which often underpin many web services, such as recommendation and content moderation. However, the raw Web data often contain spurious correlations and social biases, an…

Image ClassificationImage Editing

Discover and Mitigate Multiple Biased Subgroups in Image Classifiers

2024-03-19 · CVPR 2024 1 · Zeliang Zhang, Mingqian Feng, Zhiheng Li, Chenliang Xu

Machine learning models can perform well on in-distribution data but often fail on biased subgroups that are underrepresented in the training data, hindering the robustness of models for reliable applications. Such subgr…

Dimensionality ReductionSubgroup Discovery

Discover and Mitigate Unknown Biases with Debiasing Alternate Networks

2022-07-20 · Zhiheng Li, Anthony Hoogs, Chenliang Xu

Deep image classifiers have been found to learn biases from datasets. To mitigate the biases, most previous methods require labels of protected attributes (e.g., age, skin tone) as full-supervision, which has two limitat…

Action RecognitionFacial Attribute ClassificationOut-of-Distribution Generalization

Exposing Hidden Biases in Text-to-Image Models via Automated Prompt Search

2025-12-09 · Manos Plitsis, Giorgos Bouritsas, Vassilis Katsouros, Yannis Panagakis arxiv

Text-to-image (TTI) diffusion models have achieved remarkable visual quality, yet they have been repeatedly shown to exhibit social biases across sensitive attributes such as gender, race and age. To mitigate these biase…

Mitigating attribute amplification in counterfactual image generation

2024-03-14 · Tian Xia, Mélanie Roschewitz, Fabio De Sousa Ribeiro, Charles Jones 외

Causal generative modelling is gaining interest in medical imaging due to its ability to answer interventional and counterfactual queries. Most work focuses on generating counterfactual images that look plausible, using …

AttributecounterfactualImage Generation