{"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/normalized-contrastive-learning-for-text","title":"Normalized Contrastive Learning for Text-Video Retrieval","arxiv_id":"2212.11790","date":"2022-11-30","proceeding":null,"authors":["Yookoon Park","Mahmoud Azab","Bo Xiong","Seungwhan Moon","Florian Metze","Gourab Kundu","Kirmani Ahmed"],"abstract":"Cross-modal contrastive learning has led the recent advances in multimodal retrieval with its simplicity and effectiveness. In this work, however, we reveal that cross-modal contrastive learning suffers from incorrect normalization of the sum retrieval probabilities of each text or video instance. Specifically, we show that many test instances are either over- or under-represented during retrieval, significantly hurting the retrieval performance. To address this problem, we propose Normalized Contrastive Learning (NCL) which utilizes the Sinkhorn-Knopp algorithm to compute the instance-wise biases that properly normalize the sum retrieval probabilities of each instance so that every text and video instance is fairly represented during cross-modal retrieval. Empirical study shows that NCL brings consistent and significant gains in text-video retrieval on different model architectures, with new state-of-the-art multimodal retrieval metrics on the ActivityNet, MSVD, and MSR-VTT datasets without any architecture engineering.","url_abs":"https://arxiv.org/abs/2212.11790v1","url_pdf":"https://arxiv.org/pdf/2212.11790v1.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":"normalized-contrastive-learning-for-text","repo_url":"https://github.com/ziyang412/ucofia","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"contrastive-learning","task_name":"Contrastive Learning"},{"task_slug":"cross-modal-retrieval","task_name":"Cross-Modal Retrieval"},{"task_slug":"retrieval","task_name":"Retrieval"},{"task_slug":"video-retrieval","task_name":"Video Retrieval"}],"methods":[{"method_slug":"contrastive-learning","method_name":"Contrastive Learning"},{"method_slug":"ncl","method_name":"NCL"},{"method_slug":"test","method_name":"Test"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2212.11790","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}