{"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/attributes2classname-a-discriminative-model","title":"Attributes2Classname: A discriminative model for attribute-based unsupervised zero-shot learning","arxiv_id":"1705.01734","date":"2017-05-04","proceeding":"ICCV 2017 10","authors":["Berkan Demirel","Ramazan Gokberk Cinbis","Nazli Ikizler-Cinbis"],"abstract":"We propose a novel approach for unsupervised zero-shot learning (ZSL) of\nclasses based on their names. Most existing unsupervised ZSL methods aim to\nlearn a model for directly comparing image features and class names. However,\nthis proves to be a difficult task due to dominance of non-visual semantics in\nunderlying vector-space embeddings of class names. To address this issue, we\ndiscriminatively learn a word representation such that the similarities between\nclass and combination of attribute names fall in line with the visual\nsimilarity. Contrary to the traditional zero-shot learning approaches that are\nbuilt upon attribute presence, our approach bypasses the laborious\nattribute-class relation annotations for unseen classes. In addition, our\nproposed approach renders text-only training possible, hence, the training can\nbe augmented without the need to collect additional image data. The\nexperimental results show that our method yields state-of-the-art results for\nunsupervised ZSL in three benchmark datasets.","url_abs":"http://arxiv.org/abs/1705.01734v2","url_pdf":"http://arxiv.org/pdf/1705.01734v2.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":"attributes2classname-a-discriminative-model","repo_url":"https://github.com/berkandemirel/attributes2classname","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"zero-shot-learning","task_name":"Zero-Shot Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1705.01734","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}