Application-driven Privacy-preserving Data Publishing with Correlated\n Attributes Article Swipe

MAKE IMPACT

Aria Rezaei , Chaowei Xiao , Jie Gao , Bo Li , Sirajum Munir · YOU? · · 2018 · Open Access · · DOI: https://doi.org/10.48550/arxiv.1812.10193

Recent advances in computing have allowed for the possibility to collect\nlarge amounts of data on personal activities and private living spaces. To\naddress the privacy concerns of users in this environment, we propose a novel\nframework called PR-GAN that offers privacy-preserving mechanism using\ngenerative adversarial networks. Given a target application, PR-GAN\nautomatically modifies the data to hide sensitive attributes -- which may be\nhidden and can be inferred by machine learning algorithms -- while preserving\nthe data utility in the target application. Unlike prior works, the public's\npossible knowledge of the correlation between the target application and\nsensitive attributes is built into our modeling. We formulate our problem as an\noptimization problem, show that an optimal solution exists and use generative\nadversarial networks (GAN) to create perturbations. We further show that our\nmethod provides privacy guarantees under the Pufferfish framework, an elegant\ngeneralization of the differential privacy that allows for the modeling of\nprior knowledge on data and correlations. Through experiments, we show that our\nmethod outperforms conventional methods in effectively hiding the sensitive\nattributes while guaranteeing high performance in the target application, for\nboth property inference and training purposes. Finally, we demonstrate through\nfurther experiments that once our model learns a privacy-preserving task, such\nas hiding subjects' identity, on a group of individuals, it can perform the\nsame task on a separate group with minimal performance drops.\n

Type

preprint

Landing Page

http://arxiv.org/abs/1812.10193

PDF

https://arxiv.org/pdf/1812.10193

OA Status

green

Cited By

1

Related Works

10

OpenAlex ID

https://openalex.org/W4363681228

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W4363681228

Canonical identifier for this work in OpenAlex

DOI

https://doi.org/10.48550/arxiv.1812.10193

Digital Object Identifier

Title

Application-driven Privacy-preserving Data Publishing with Correlated\n Attributes

Work title

Type

preprint

OpenAlex work type

Publication year

2018

Year of publication

Publication date

2018-12-25

Full publication date if available

Authors

Aria Rezaei, Chaowei Xiao, Jie Gao, Bo Li, Sirajum Munir

List of authors in order

Landing page

https://arxiv.org/abs/1812.10193

Publisher landing page

PDF URL

https://arxiv.org/pdf/1812.10193

Direct link to full text PDF

Open access

Yes

Whether a free full text is available

OA status

green

Open access status per OpenAlex

OA URL

https://arxiv.org/pdf/1812.10193

Direct OA link when available

Concepts

Computer science, Differential privacy, Inference, Adversarial system, Task (project management), Generalization, Data publishing, Generative grammar, Machine learning, Generative model, Property (philosophy), Artificial intelligence, Identity (music), Data mining, Theoretical computer science, Publishing, Mathematics, Mathematical analysis, Physics, Law, Economics, Epistemology, Philosophy, Management, Political science, Acoustics

Top concepts (fields/topics) attached by OpenAlex

Cited by

1

Total citation count in OpenAlex

Citations by year (recent)

2023: 1

Per-year citation counts (last 5 years)

Related works (count)

10

Other works algorithmically related by OpenAlex

Full payload