{"title":"信息泄漏公理法的一般化","authors":"Mohammad Amin Zarrabian, Parastoo Sadeghi","doi":"arxiv-2409.04108","DOIUrl":null,"url":null,"abstract":"In this paper, we extend the framework of quantitative information flow (QIF)\nto include adversaries that use Kolmogorov-Nagumo $f$-mean to infer secrets of\na private system. Specifically, in our setting, an adversary uses\nKolmogorov-Nagumo $f$-mean to compute its best actions before and after\nobserving the system's randomized outputs. This leads to generalized notions of\nprior and posterior vulnerability and generalized axiomatic relations that we\nwill derive to elucidate how these $f$-mean based vulnerabilities interact with\neach other. We demonstrate usefulness of this framework by showing how some\nnotions of leakage that had been derived outside of the QIF framework and so\nfar seemed incompatible with it are indeed explainable via such extension of\nQIF. These leakage measures include $\\alpha$-leakage, which is the same as\nArimoto mutual information of order $\\alpha$, maximal $\\alpha$-leakage which is\nthe $\\alpha$-leakage capacity, and $(\\alpha,\\beta)$ leakage, which is a\ngeneralization of the above and captures local differential privacy as a\nspecial case. We also propose a new pointwise notion of gain function, which we\ncoin pointwise information gain. We show that this pointwise information gain\ncan explain R\\'eyni divergence and Sibson mutual information of order $\\alpha\n\\in [0,\\infty]$ as the Kolmogorov-Nagumo average of the gain with a proper\nchoice of function $f$.","PeriodicalId":501082,"journal":{"name":"arXiv - MATH - Information Theory","volume":"37 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-09-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A Generalization of Axiomatic Approach to Information Leakage\",\"authors\":\"Mohammad Amin Zarrabian, Parastoo Sadeghi\",\"doi\":\"arxiv-2409.04108\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In this paper, we extend the framework of quantitative information flow (QIF)\\nto include adversaries that use Kolmogorov-Nagumo $f$-mean to infer secrets of\\na private system. Specifically, in our setting, an adversary uses\\nKolmogorov-Nagumo $f$-mean to compute its best actions before and after\\nobserving the system's randomized outputs. This leads to generalized notions of\\nprior and posterior vulnerability and generalized axiomatic relations that we\\nwill derive to elucidate how these $f$-mean based vulnerabilities interact with\\neach other. We demonstrate usefulness of this framework by showing how some\\nnotions of leakage that had been derived outside of the QIF framework and so\\nfar seemed incompatible with it are indeed explainable via such extension of\\nQIF. These leakage measures include $\\\\alpha$-leakage, which is the same as\\nArimoto mutual information of order $\\\\alpha$, maximal $\\\\alpha$-leakage which is\\nthe $\\\\alpha$-leakage capacity, and $(\\\\alpha,\\\\beta)$ leakage, which is a\\ngeneralization of the above and captures local differential privacy as a\\nspecial case. We also propose a new pointwise notion of gain function, which we\\ncoin pointwise information gain. We show that this pointwise information gain\\ncan explain R\\\\'eyni divergence and Sibson mutual information of order $\\\\alpha\\n\\\\in [0,\\\\infty]$ as the Kolmogorov-Nagumo average of the gain with a proper\\nchoice of function $f$.\",\"PeriodicalId\":501082,\"journal\":{\"name\":\"arXiv - MATH - Information Theory\",\"volume\":\"37 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-06\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - MATH - Information Theory\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.04108\",\"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 - MATH - Information Theory","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.04108","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
A Generalization of Axiomatic Approach to Information Leakage
In this paper, we extend the framework of quantitative information flow (QIF)
to include adversaries that use Kolmogorov-Nagumo $f$-mean to infer secrets of
a private system. Specifically, in our setting, an adversary uses
Kolmogorov-Nagumo $f$-mean to compute its best actions before and after
observing the system's randomized outputs. This leads to generalized notions of
prior and posterior vulnerability and generalized axiomatic relations that we
will derive to elucidate how these $f$-mean based vulnerabilities interact with
each other. We demonstrate usefulness of this framework by showing how some
notions of leakage that had been derived outside of the QIF framework and so
far seemed incompatible with it are indeed explainable via such extension of
QIF. These leakage measures include $\alpha$-leakage, which is the same as
Arimoto mutual information of order $\alpha$, maximal $\alpha$-leakage which is
the $\alpha$-leakage capacity, and $(\alpha,\beta)$ leakage, which is a
generalization of the above and captures local differential privacy as a
special case. We also propose a new pointwise notion of gain function, which we
coin pointwise information gain. We show that this pointwise information gain
can explain R\'eyni divergence and Sibson mutual information of order $\alpha
\in [0,\infty]$ as the Kolmogorov-Nagumo average of the gain with a proper
choice of function $f$.