Certified Data Removal from Machine Learning Models

被引：0

作者：

Guo, Chuan ^{[1
]}

Goldstein, Tom ^{[2
]}

Hannun, Awni ^{[2
]}

van der Maaten, Laurens ^{[2
]}

机构：

[1] Cornell Univ, Dept Comp Sci, Ithaca, NY 14850 USA

[2] Facebook AI Res, New York, NY 10003 USA

来源：

INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 119 | 2020年 / 119卷

关键词：

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Good data stewardship requires removal of data at the request of the data's owner. This raises the question if and how a trained machine-learning model, which implicitly stores information about its training data, should be affected by such a removal request. Is it possible to "remove" data from a machine-learning model? We study this problem by defining certified removal: a very strong theoretical guarantee that a model from which data is removed cannot be distinguished from a model that never observed the data to begin with. We develop a certified-removal mechanism for linear classifiers and empirically study learning settings in which this mechanism is practical.

引用

页数：11

共 50 条

[1] Certified Data Removal from Machine Learning Models
Guo, Chuan
Goldstein, Tom
Hannun, Awni
van der Maaten, Laurens
[J]. 25TH AMERICAS CONFERENCE ON INFORMATION SYSTEMS (AMCIS 2019), 2019,
[2] Certified Machine-Learning Models
Damiani, Ernesto
Ardagna, Claudio A.
[J]. SOFSEM 2020: THEORY AND PRACTICE OF COMPUTER SCIENCE, 2020, 12011 : 3 - 15
[3] Approximate Data Deletion from Machine Learning Models
Izzo, Zachary
Smart, Mary Anne
Chaudhuri, Kamalika
Zou, James
[J]. 24TH INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS (AISTATS), 2021, 130
[4] Learning EPON delay models from data: a machine learning approach
Alberto Hernandez, Jose
Ebrahimzadeh, Amin
Maier, Martin
Larrabeiti, David
[J]. JOURNAL OF OPTICAL COMMUNICATIONS AND NETWORKING, 2021, 13 (12) : 322 - 330
[5] Generating models of mental retardation from data with machine learning
Mani, S
McDermott, S
Pazzani, MJ
[J]. 1997 IEEE KNOWLEDGE AND DATA ENGINEERING EXCHANGE WORKSHOP, PROCEEDINGS, 1997, : 114 - 119
[6] Information Leakage from Data Updates in Machine Learning Models
Hui, Tian
Farokhi, Farhad
Ohrimenko, Olga
[J]. PROCEEDINGS OF THE 16TH ACM WORKSHOP ON ARTIFICIAL INTELLIGENCE AND SECURITY, AISEC 2023, 2023, : 35 - 41
[7] Towards Machine Learning of Predictive Models from Ecological Data
Tamaddoni-Nezhad, Alireza
Bohan, David
Raybould, Alan
Muggleton, Stephen
[J]. INDUCTIVE LOGIC PROGRAMMING, ILP 2014, 2015, 9046 : 154 - 167
[8] Stealing Your Data from Compressed Machine Learning Models
Xu, Nuo
Liu, Qi
Liu, Tao
Liu, Zihao
Guo, Xiaochen
Wen, Wujie
[J]. PROCEEDINGS OF THE 2020 57TH ACM/EDAC/IEEE DESIGN AUTOMATION CONFERENCE (DAC), 2020,
[9] Learning earth system models from observations: machine learning or data assimilation?
Geer, A. J.
[J]. PHILOSOPHICAL TRANSACTIONS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES, 2021, 379 (2194):
[10] Improving the accuracy of machine-learning models with data from machine test repetitions
Andres Bustillo
Roberto Reis
Alisson R. Machado
Danil Yu. Pimenov
[J]. Journal of Intelligent Manufacturing, 2022, 33 : 203 - 221

← 1 2 3 4 5 →