{"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/the-bosaris-toolkit-theory-algorithms-and","title":"The BOSARIS Toolkit: Theory, Algorithms and Code for Surviving the New DCF","arxiv_id":"1304.2865","date":"2013-04-10","proceeding":null,"authors":["Niko Brümmer","Edward de Villiers"],"abstract":"The change of two orders of magnitude in the 'new DCF' of NIST's SRE'10,\nrelative to the 'old DCF' evaluation criterion, posed a difficult challenge for\nparticipants and evaluator alike. Initially, participants were at a loss as to\nhow to calibrate their systems, while the evaluator underestimated the required\nnumber of evaluation trials. After the fact, it is now obvious that both\ncalibration and evaluation require very large sets of trials. This poses the\nchallenges of (i) how to decide what number of trials is enough, and (ii) how\nto process such large data sets with reasonable memory and CPU requirements.\nAfter SRE'10, at the BOSARIS Workshop, we built solutions to these problems\ninto the freely available BOSARIS Toolkit. This paper explains the principles\nand algorithms behind this toolkit. The main contributions of the toolkit are:\n1. The Normalized Bayes Error-Rate Plot, which analyses likelihood- ratio\ncalibration over a wide range of DCF operating points. These plots also help in\njudging the adequacy of the sizes of calibration and evaluation databases. 2.\nEfficient algorithms to compute DCF and minDCF for large score files, over the\nrange of operating points required by these plots. 3. A new score file format,\nwhich facilitates working with very large trial lists. 4. A faster logistic\nregression optimizer for fusion and calibration. 5. A principled way to define\nEER (equal error rate), which is of practical interest when the absolute error\ncount is small.","url_abs":"http://arxiv.org/abs/1304.2865v1","url_pdf":"http://arxiv.org/pdf/1304.2865v1.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":"the-bosaris-toolkit-theory-algorithms-and","repo_url":"https://github.com/bsxfan/PYLLR","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":null,"task_name":"CPU"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}