Presentation 2019-05-17
Examination of quantitative evaluation of attack method to restore the anonymized data
Takaya Yamazoe, Kazumasa Omote,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) With the increasing scale of data, integrated analysis of data distributed to multiple institutions is required. On the other hand, collecting data from multiple institutions often involves privacy risks, and there is a methods called anonymization as a representative measure. In this research, we improve Model Inversion Attack, which is one of attacks against models, and propose the attack method against anonymized data which we call Data Inversion Attack. In the evaluation, the data that has been pk anonymized and matrix decomposition anonymized that can handle high-dimensional data is targeted, and the restoration of the original data is evaluated quantitatively. In particular, using the distribution of the original data as a baseline, we have shown that our attack method can contribute to restoration of anonymized data with higher accuracy than the baseline.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) anonymization / privacy / restore / Inversion Attack
Paper # ISEC2019-4
Date of Issue 2019-05-10 (ISEC)

Conference Information
Committee ISEC
Conference Date 2019/5/17(1days)
Place (in Japanese) (See Japanese page)
Place (in English) Kikai-Shinko-Kaikan Bldg.
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Atsushi Fujioka(Kanagawa Univ.)
Vice Chair Shiho Moriai(NICT) / Shoichi Hirose(Univ. of Fukui)
Secretary Shiho Moriai(Tokai Univ.) / Shoichi Hirose(NICT)
Assistant Kazunari Omote(Tsukuba Univ.) / Yuuji Suga(IIJ)

Paper Information
Registration To Technical Committee on Information Security
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Examination of quantitative evaluation of attack method to restore the anonymized data
Sub Title (in English)
Keyword(1) anonymization
Keyword(2) privacy
Keyword(3) restore
Keyword(4) Inversion Attack
1st Author's Name Takaya Yamazoe
1st Author's Affiliation Tsukuba University(Tsukuba Univ.)
2nd Author's Name Kazumasa Omote
2nd Author's Affiliation Tsukuba University(Tsukuba Univ.)
Date 2019-05-17
Paper # ISEC2019-4
Volume (vol) vol.119
Number (no) ISEC-40
Page pp.pp.7-13(ISEC),
#Pages 7
Date of Issue 2019-05-10 (ISEC)