{"title":"光滑巴拿赫空间中的经验伯恩斯坦","authors":"Diego Martinez-Taboada, Aaditya Ramdas","doi":"arxiv-2409.06060","DOIUrl":null,"url":null,"abstract":"Existing concentration bounds for bounded vector-valued random variables\ninclude extensions of the scalar Hoeffding and Bernstein inequalities. While\nthe latter is typically tighter, it requires knowing a bound on the variance of\nthe random variables. We derive a new vector-valued empirical Bernstein\ninequality, which makes use of an empirical estimator of the variance instead\nof the true variance. The bound holds in 2-smooth separable Banach spaces,\nwhich include finite dimensional Euclidean spaces and separable Hilbert spaces.\nThe resulting confidence sets are instantiated for both the batch setting\n(where the sample size is fixed) and the sequential setting (where the sample\nsize is a stopping time). The confidence set width asymptotically exactly\nmatches that achieved by Bernstein in the leading term. The method and\nsupermartingale proof technique combine several tools of Pinelis (1994) and\nWaudby-Smith and Ramdas (2024).","PeriodicalId":501379,"journal":{"name":"arXiv - STAT - Statistics Theory","volume":"8 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Empirical Bernstein in smooth Banach spaces\",\"authors\":\"Diego Martinez-Taboada, Aaditya Ramdas\",\"doi\":\"arxiv-2409.06060\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Existing concentration bounds for bounded vector-valued random variables\\ninclude extensions of the scalar Hoeffding and Bernstein inequalities. While\\nthe latter is typically tighter, it requires knowing a bound on the variance of\\nthe random variables. We derive a new vector-valued empirical Bernstein\\ninequality, which makes use of an empirical estimator of the variance instead\\nof the true variance. The bound holds in 2-smooth separable Banach spaces,\\nwhich include finite dimensional Euclidean spaces and separable Hilbert spaces.\\nThe resulting confidence sets are instantiated for both the batch setting\\n(where the sample size is fixed) and the sequential setting (where the sample\\nsize is a stopping time). The confidence set width asymptotically exactly\\nmatches that achieved by Bernstein in the leading term. The method and\\nsupermartingale proof technique combine several tools of Pinelis (1994) and\\nWaudby-Smith and Ramdas (2024).\",\"PeriodicalId\":501379,\"journal\":{\"name\":\"arXiv - STAT - Statistics Theory\",\"volume\":\"8 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-09\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - STAT - Statistics Theory\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.06060\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - STAT - Statistics Theory","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.06060","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Existing concentration bounds for bounded vector-valued random variables
include extensions of the scalar Hoeffding and Bernstein inequalities. While
the latter is typically tighter, it requires knowing a bound on the variance of
the random variables. We derive a new vector-valued empirical Bernstein
inequality, which makes use of an empirical estimator of the variance instead
of the true variance. The bound holds in 2-smooth separable Banach spaces,
which include finite dimensional Euclidean spaces and separable Hilbert spaces.
The resulting confidence sets are instantiated for both the batch setting
(where the sample size is fixed) and the sequential setting (where the sample
size is a stopping time). The confidence set width asymptotically exactly
matches that achieved by Bernstein in the leading term. The method and
supermartingale proof technique combine several tools of Pinelis (1994) and
Waudby-Smith and Ramdas (2024).