[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83876-en":3,"doc-seo-83876-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},83876,8796095461564,"Liam","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","Virtual Category Guided Continual Generalized Category Discovery","Continual Generalized Category Discovery (C-GCD) incrementally identifies novel categories from sequential unlabeled data while preserving recognition of previously known classes, enabling open-world visual learning. The key bottleneck is ambiguity: unlabeled samples cannot be confidently assigned to known classes nor reliably grouped as novel ones, making pseudo-labeling unstable and biased toward familiar categories. The work proposes Virtual Category-Guided C-GCD, adapting Virtual Category Learning to assign uncertain samples to temporary virtual categories, improving safe learning and reducing prediction bias. Expanded Neighborhood Contrastive Learning further stabilizes discovery across sessions by enhancing representation separability for old and emerging classes. Experiments on CIFAR-100, Tiny ImageNet, and ImageNet-100 show consistent state-of-the-art gains.","arXiv :2607 .04984v 1 [ cs .CV] 6 Jul 2026  \nVirtual Category-Guided Continual Generalized Category Discovery  \nJiahui Xiong 1 , Qiuxia Lai2 *, and Hongsong Wang 1 ,3 *  \n1 School of Computer Science and Engineering, Southeast University, Nanjing 210096, China  \n2 State Key Laboratory of Media Convergence and Communication, Communication University of China, Beijing 100024, China  \n3 Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications (Southeast University), Ministry of Education, China  \n{hongsongwang, [jiahuixiong}@seu.edu.cn](jiahuixiong}@seu.edu.cn) , [qxlai@cuc.edu.cn](qxlai@cuc.edu.cn)  \nAbstract. Continual Generalized Category Discovery (C-GCD) aims to incrementally identify novel categories from sequential unlabeled data while preserving recognition of known classes, which is an essential capability for open-world visual learning. A major bottleneck lies in ambiguous unlabeled samples that cannot be confidently assigned to known classes nor reliably grouped as novel ones, making pseudo-labeling brittle and often biasing learning toward familiar categories. In this work, we introduce Virtual Category-Guided Continual Generalized Category Discovery by adapting Virtual Category Learning (VCL) to the continual setting. Our method identifies uncertain samples and assigns them to temporary virtual categories, enabling safe and informative learning from unlabeled streams without injecting noisy labels, while improving unlabeled data utilization and mitigating prediction bias. To further stabilize discovery across sessions and enhance class separation, we augment VCL with Expanded Neighborhood Contrastive Learning (ENCL), which exploits extended neighborhood relations and an adaptive margin to learn more discriminative and well-separated representations for both old and emerging classes. Extensive experiments on CIFAR-100, Tiny ImageNet, and ImageNet-100 demonstrate that our approach consistently outperforms state-of-the-art methods, establishing a scalable and effective solution for C-GCD. Code is on: [https://github.com/Mrxjh105/VC-CGCD](https://github.com/Mrxjh105/VC-CGCD)  \nKeywords: Continual Generalized Category Discovery · Continual Learning · Open-World Recognition  \n1 Introduction  \nModern visual recognition has made striking progress in closed-world settings where the label space is fixed and exhaustively defined [8, 22] . However, real de  \nployments, such as home robots, autonomous driving, and large-scale monitoring,∗ Corresponding authors.  \n2 J. Xiong et al.  \noperate in open and evolving environments, where novel categories appear continually and must be incorporated without retraining from scratch [9, 24] . This requirement exposes a fundamental challenge for long-lived learners: they must expand their category set over time while retaining competence on previously learned classes, i.e., the stability–plasticity dilemma. When naively fine-tuned on new data, deep networks are prone to catastrophic forgetting [14,16,20], making continual open-world recognition particularly unstable.  \nTo address emerging categories, Novel Category Discovery (NCD) [7] studies how to discover unseen classes from unlabeled data given labeled known classes, typically under the assumption that the unlabeled pool contains only novel categories. This assumption rarely holds in realistic streams where old and new categories naturally co-exist. Generalized Category Discovery (GCD) [23] relaxes the setting by allowing unlabeled data to mix known and novel classes, requiring simultaneous recognition of known instances and clustering of novel ones. However, most GCD methods remain batch-oriented and assume access to the full unlabeled pool, making them ill-suited to continuous data arrival. Continual Generalized Category Discovery (C-GCD) [30] pushes category discovery into a truly continual regime: after an offline phase trained on labeled known classes, the model receives a sequence o","cbCaikU8Mm2Fs6RY","https://ap.wps.com/l/cbCaikU8Mm2Fs6RY","pdf",4896908,3,1,24,"English","en",105,"# Introduction\n## Problem setting and challenges\n## Related work: NCD, GCD, and C-GCD\n## Ambiguity and risks of pseudo-labeling\n## Main idea: virtual categories\n## Method overview: VCL adaptation and ENCL","[{\"question\":\"What problem does Continual Generalized Category Discovery (C-GCD) address?\",\"answer\":\"C-GCD targets learning new categories from a stream of unlabeled sessions while retaining performance on previously known classes, without retraining from scratch. It is designed for open-world visual learning where categories evolve over time.\"},{\"question\":\"Why is pseudo-labeling difficult in C-GCD?\",\"answer\":\"Unlabeled samples can be ambiguous, meaning they are neither confidently known nor reliably novel. Hard pseudo-labeling or aggressive clustering can introduce persistent label noise and confirmation bias that accumulates across sessions.\"},{\"question\":\"How does the proposed Virtual Category-Guided approach improve learning?\",\"answer\":\"It assigns uncertain samples to temporary virtual categories, acting as a controlled buffer that absorbs informative structure from ambiguous data. This avoids injecting noisy hard labels into the known/novel split and mitigates prediction bias.\"}]",1784191166,60,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"virtual-category-guided-continual-generalized-category-discovery","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/virtual-category-guided-continual-generalized-category-discovery/83876/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-26","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does Continual Generalized Category Discovery (C-GCD) address?","Question",{"text":75,"@type":76},"C-GCD targets learning new categories from a stream of unlabeled sessions while retaining performance on previously known classes, without retraining from scratch. It is designed for open-world visual learning where categories evolve over time.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why is pseudo-labeling difficult in C-GCD?",{"text":80,"@type":76},"Unlabeled samples can be ambiguous, meaning they are neither confidently known nor reliably novel. Hard pseudo-labeling or aggressive clustering can introduce persistent label noise and confirmation bias that accumulates across sessions.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the proposed Virtual Category-Guided approach improve learning?",{"text":84,"@type":76},"It assigns uncertain samples to temporary virtual categories, acting as a controlled buffer that absorbs informative structure from ambiguous data. This avoids injecting noisy hard labels into the known/novel split and mitigates prediction bias.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,109,114,119,122,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":29,"slug":108},5,"Comic","comic",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]