{"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/attention-based-ensemble-for-deep-metric","title":"Attention-based Ensemble for Deep Metric Learning","arxiv_id":"1804.00382","date":"2018-04-02","proceeding":"ECCV 2018 9","authors":["Wonsik Kim","Bhavya Goyal","Kunal Chawla","Jungmin Lee","Keunjoo Kwon"],"abstract":"Deep metric learning aims to learn an embedding function, modeled as deep\nneural network. This embedding function usually puts semantically similar\nimages close while dissimilar images far from each other in the learned\nembedding space. Recently, ensemble has been applied to deep metric learning to\nyield state-of-the-art results. As one important aspect of ensemble, the\nlearners should be diverse in their feature embeddings. To this end, we propose\nan attention-based ensemble, which uses multiple attention masks, so that each\nlearner can attend to different parts of the object. We also propose a\ndivergence loss, which encourages diversity among the learners. The proposed\nmethod is applied to the standard benchmarks of deep metric learning and\nexperimental results show that it outperforms the state-of-the-art methods by a\nsignificant margin on image retrieval tasks.","url_abs":"http://arxiv.org/abs/1804.00382v2","url_pdf":"http://arxiv.org/pdf/1804.00382v2.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":[],"tasks":[{"task_slug":"diversity","task_name":"Diversity"},{"task_slug":"image-retrieval","task_name":"Image Retrieval"},{"task_slug":"metric-learning","task_name":"Metric Learning"},{"task_slug":"retrieval","task_name":"Retrieval"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-retrieval-on-in-shop","task":"Image Retrieval","dataset":"In-Shop","model":"ABE-8","rank_in_archive_order":7,"of":7,"metrics":{"R@1":"87.3"},"uses_additional_data":false},{"leaderboard":"/sota/image-retrieval-on-sop","task":"Image Retrieval","dataset":"SOP","model":"ABE-8","rank_in_archive_order":12,"of":14,"metrics":{"R@1":"76.3"},"uses_additional_data":false},{"leaderboard":"/sota/metric-learning-on-cars196","task":"Metric Learning","dataset":"CARS196","model":"ABE-8-512","rank_in_archive_order":27,"of":36,"metrics":{"R@1":"85.2"},"uses_additional_data":true},{"leaderboard":"/sota/metric-learning-on-cub-200-2011","task":"Metric Learning","dataset":"CUB-200-2011","model":"ABE-8-512","rank_in_archive_order":26,"of":30,"metrics":{"R@1":"60.6"},"uses_additional_data":true}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.00382","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}