{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/matched-sample-selection-with-gans-for","title":"Matched sample selection with GANs for mitigating attribute confounding","arxiv_id":"2103.13455","date":"2021-03-24","proceeding":null,"authors":["Chandan Singh","Guha Balakrishnan","Pietro Perona"],"abstract":"Measuring biases of vision systems with respect to protected attributes like gender and age is critical as these systems gain widespread use in society. However, significant correlations between attributes in benchmark datasets make it difficult to separate algorithmic bias from dataset bias. To mitigate such attribute confounding during bias analysis, we propose a matching approach that selects a subset of images from the full dataset with balanced attribute distributions across protected attributes. Our matching approach first projects real images onto a generative adversarial network (GAN)'s latent space in a manner that preserves semantic attributes. It then finds image matches in this latent space across a chosen protected attribute, yielding a dataset where semantic and perceptual attributes are balanced across the protected attribute. We validate projection and matching strategies with qualitative, quantitative, and human annotation experiments. We demonstrate our work in the context of gender bias in multiple open-source facial-recognition classifiers and find that bias persists after removing key confounders via matching. Code and documentation to reproduce the results here and apply the methods to new data is available at https://github.com/csinva/matching-with-gans .","url_abs":"https://arxiv.org/abs/2103.13455v1","url_pdf":"https://arxiv.org/pdf/2103.13455v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"matched-sample-selection-with-gans-for","repo_url":"https://github.com/csinva/matching-with-gans","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"face-generation","task_name":"Face Generation"},{"task_slug":"gender-bias-detection","task_name":"Gender Bias Detection"},{"task_slug":null,"task_name":"Generative Adversarial Network"},{"task_slug":"selection-bias","task_name":"Selection bias"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"path-length-regularization","method_name":"Path Length Regularization"},{"method_slug":"r1-regularization","method_name":"R1 Regularization"},{"method_slug":"stylegan2","method_name":"StyleGAN2"},{"method_slug":"weight-demodulation","method_name":"Weight Demodulation"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2103.13455","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}