{"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/imbalanced-deep-learning-by-minority-class","title":"Imbalanced Deep Learning by Minority Class Incremental Rectification","arxiv_id":"1804.10851","date":"2018-04-28","proceeding":null,"authors":["Qi Dong","Shaogang Gong","Xiatian Zhu"],"abstract":"Model learning from class imbalanced training data is a long-standing and\nsignificant challenge for machine learning. In particular, existing deep\nlearning methods consider mostly either class balanced data or moderately\nimbalanced data in model training, and ignore the challenge of learning from\nsignificantly imbalanced training data. To address this problem, we formulate a\nclass imbalanced deep learning model based on batch-wise incremental minority\n(sparsely sampled) class rectification by hard sample mining in majority\n(frequently sampled) classes during model training. This model is designed to\nminimise the dominant effect of majority classes by discovering sparsely\nsampled boundaries of minority classes in an iterative batch-wise learning\nprocess. To that end, we introduce a Class Rectification Loss (CRL) function\nthat can be deployed readily in deep network architectures. Extensive\nexperimental evaluations are conducted on three imbalanced person attribute\nbenchmark datasets (CelebA, X-Domain, DeepFashion) and one balanced object\ncategory benchmark dataset (CIFAR-100). These experimental results demonstrate\nthe performance advantages and model scalability of the proposed batch-wise\nincremental minority class rectification model over the existing\nstate-of-the-art models for addressing the problem of imbalanced data learning.","url_abs":"http://arxiv.org/abs/1804.10851v1","url_pdf":"http://arxiv.org/pdf/1804.10851v1.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":"imbalanced-deep-learning-by-minority-class","repo_url":"https://github.com/AemikaChow/DATASOURCE","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"deep-learning","task_name":"Deep Learning"},{"task_slug":"facial-attribute-classification","task_name":"Facial Attribute Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.10851","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}