{"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/twin-contrastive-learning-with-noisy-labels","title":"Twin Contrastive Learning with Noisy Labels","arxiv_id":"2303.06930","date":"2023-03-13","proceeding":"CVPR 2023 1","authors":["Zhizhong Huang","Junping Zhang","Hongming Shan"],"abstract":"Learning from noisy data is a challenging task that significantly degenerates the model performance. In this paper, we present TCL, a novel twin contrastive learning model to learn robust representations and handle noisy labels for classification. Specifically, we construct a Gaussian mixture model (GMM) over the representations by injecting the supervised model predictions into GMM to link label-free latent variables in GMM with label-noisy annotations. Then, TCL detects the examples with wrong labels as the out-of-distribution examples by another two-component GMM, taking into account the data distribution. We further propose a cross-supervision with an entropy regularization loss that bootstraps the true targets from model predictions to handle the noisy labels. As a result, TCL can learn discriminative representations aligned with estimated labels through mixup and contrastive learning. Extensive experimental results on several standard benchmarks and real-world datasets demonstrate the superior performance of TCL. In particular, TCL achieves 7.5\\% improvements on CIFAR-10 with 90\\% noisy label -- an extremely noisy scenario. The source code is available at \\url{https://github.com/Hzzone/TCL}.","url_abs":"https://arxiv.org/abs/2303.06930v1","url_pdf":"https://arxiv.org/pdf/2303.06930v1.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":"twin-contrastive-learning-with-noisy-labels","repo_url":"https://github.com/hzzone/tcl","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"contrastive-learning","task_name":"Contrastive Learning"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"learning-with-noisy-labels","task_name":"Learning with noisy labels"}],"methods":[{"method_slug":"contrastive-learning","method_name":"Contrastive Learning"},{"method_slug":"entropy-regularization","method_name":"Entropy Regularization"},{"method_slug":"mixup","method_name":"Mixup"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-classification-on-mini-webvision-1-0","task":"Image Classification","dataset":"mini WebVision 1.0","model":"TCL","rank_in_archive_order":20,"of":47,"metrics":{"ImageNet Top-1 Accuracy":"75.4","ImageNet Top-5 Accuracy":"92.4","Top-1 Accuracy":"79.1","Top-5 Accuracy":"92.3"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2303.06930","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}