{"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/learning-deep-embeddings-with-histogram-loss","title":"Learning Deep Embeddings with Histogram Loss","arxiv_id":"1611.00822","date":"2016-11-02","proceeding":"NeurIPS 2016 12","authors":["Evgeniya Ustinova","Victor Lempitsky"],"abstract":"We suggest a loss for learning deep embeddings. The new loss does not\nintroduce parameters that need to be tuned and results in very good embeddings\nacross a range of datasets and problems. The loss is computed by estimating two\ndistribution of similarities for positive (matching) and negative\n(non-matching) sample pairs, and then computing the probability of a positive\npair to have a lower similarity score than a negative pair based on the\nestimated similarity distributions. We show that such operations can be\nperformed in a simple and piecewise-differentiable manner using 1D histograms\nwith soft assignment operations. This makes the proposed loss suitable for\nlearning deep embeddings using stochastic optimization. In the experiments, the\nnew loss performs favourably compared to recently proposed alternatives.","url_abs":"http://arxiv.org/abs/1611.00822v1","url_pdf":"http://arxiv.org/pdf/1611.00822v1.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":"learning-deep-embeddings-with-histogram-loss","repo_url":"https://github.com/madkn/HistogramLoss","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"stochastic-optimization","task_name":"Stochastic Optimization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1611.00822","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}