Adaptive Curvature Hierarchical Hyperbolic Contrastive Learning for Fine-Grained Cross-Modal Retrieval

Adaptive Curvature Hierarchical Hyperbolic Contrastive Learning for Fine-Grained Cross-Modal Retrieval | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 9 April 2026 V1 Latest version Share on Adaptive Curvature Hierarchical Hyperbolic Contrastive Learning for Fine-Grained Cross-Modal Retrieval Authors : Kexin Ruan 0009-0006-3943-7308 [email protected] and Haoyu Yan Authors Info & Affiliations https://doi.org/10.22541/au.177575360.09137481/v1 53 views 18 downloads Contents Abstract Supplementary Material Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Cross-modal retrieval struggles with modeling hierarchical structures and distinguishing hard negative samples in multi-modal data using traditional Euclidean space. This paper introduces Hierarchical Hyperbolic Contrastive Learning (HHCL), a novel framework embedding multimodal features into a shared hyperbolic space to overcome these issues. HHCL leverages multimodal encoders and hyperbolic projection layers to map features onto the Poincaré ball model. Its core innovation is an adaptive curvature hyperbolic contrastive loss, dynamically learning and optimizing curvature parameters based on local data characteristics. This captures multi-scale hierarchical information and addresses hard negative samples. Evaluated on fine-grained crossmodal retrieval tasks across MS-COCO, Flickr30K, and a CUB-200-2011 subset, HHCL consistently achieves state-of-the-art performance, significantly outperforming Euclidean baselines and fixed-curvature hyperbolic approaches. Ablation studies validate the adaptive curvature's effectiveness. Qualitative analyses demonstrate HHCL's superior ability to align fine-grained semantic details, positioning it as a robust solution for complex cross-modal matching. Supplementary Material File (hhcl.pdf) Download 2.20 MB Information & Authors Information Version history V1 Version 1 09 April 2026 Copyright This work is licensed under a Creative Commons Attribution 4.0 International License Keywords adaptive curvature contrastive learning cross-modal retrieval hierarchical structures hyperbolic space Authors Affiliations Kexin Ruan 0009-0006-3943-7308 [email protected] Xihua University View all articles by this author Haoyu Yan Xihua University View all articles by this author Metrics & Citations Metrics Article Usage 53 views 18 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Kexin Ruan, Haoyu Yan. Adaptive Curvature Hierarchical Hyperbolic Contrastive Learning for Fine-Grained Cross-Modal Retrieval. Authorea . 09 April 2026. DOI: https://doi.org/10.22541/au.177575360.09137481/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.177575360.09137481/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9fdfa938ceac4807',t:'MTc3OTE1NzYzMA=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();