{"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/deep-triplet-ranking-networks-for-one-shot","title":"Deep Triplet Ranking Networks for One-Shot Recognition","arxiv_id":"1804.07275","date":"2018-04-19","proceeding":null,"authors":["Meng Ye","Yuhong Guo"],"abstract":"Despite the breakthroughs achieved by deep learning models in conventional\nsupervised learning scenarios, their dependence on sufficient labeled training\ndata in each class prevents effective applications of these deep models in\nsituations where labeled training instances for a subset of novel classes are\nvery sparse -- in the extreme case only one instance is available for each\nclass. To tackle this natural and important challenge, one-shot learning, which\naims to exploit a set of well labeled base classes to build classifiers for the\nnew target classes that have only one observed instance per class, has recently\nreceived increasing attention from the research community. In this paper we\npropose a novel end-to-end deep triplet ranking network to perform one-shot\nlearning. The proposed approach learns class universal image embeddings on the\nwell labeled base classes under a triplet ranking loss, such that the instances\nfrom new classes can be categorized based on their similarity with the one-shot\ninstances in the learned embedding space. Moreover, our approach can naturally\nincorporate the available one-shot instances from the new classes into the\nembedding learning process to improve the triplet ranking model. We conduct\nexperiments on two popular datasets for one-shot learning. The results show the\nproposed approach achieves better performance than the state-of-the- art\ncomparison methods.","url_abs":"http://arxiv.org/abs/1804.07275v1","url_pdf":"http://arxiv.org/pdf/1804.07275v1.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":"deep-triplet-ranking-networks-for-one-shot","repo_url":"https://github.com/Mushfequr-Rahman/Omniglot_baselines","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"one-shot-learning","task_name":"One-Shot Learning"},{"task_slug":null,"task_name":"Triplet"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}