Semantic deep hiding for robust unlearnable examples
Ensuring data privacy and protection has become paramount in the era of deep learning. Unlearnable examples are proposed to mislead the deep learning models and prevent data from unauthorized exploration by adding small perturbations to data. However, such perturbations (e.g., noise, texture, color...
Saved in:
Main Authors: | , , , , , |
---|---|
Other Authors: | |
Format: | Article |
Language: | English |
Published: |
2024
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/180637 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
id |
sg-ntu-dr.10356-180637 |
---|---|
record_format |
dspace |
spelling |
sg-ntu-dr.10356-1806372024-10-15T08:14:56Z Semantic deep hiding for robust unlearnable examples Meng, Ruohan Yi, Chenyu Yu, Yi Yang, Siyuan Shen, Bingquan Kot, Alex C. School of Electrical and Electronic Engineering Rapid-Rich Object Search (ROSE) Lab Engineering Unlearnable examples Deep hiding Ensuring data privacy and protection has become paramount in the era of deep learning. Unlearnable examples are proposed to mislead the deep learning models and prevent data from unauthorized exploration by adding small perturbations to data. However, such perturbations (e.g., noise, texture, color change) predominantly impact low-level features, making them vulnerable to common countermeasures. In contrast, semantic images with intricate shapes have a wealth of high-level features, making them more resilient to countermeasures and potential for producing robust unlearnable examples. In this paper, we propose a Deep Hiding (DH) scheme that adaptively hides semantic images enriched with high-level features. We employ an Invertible Neural Network (INN) to invisibly integrate predefined images, inherently hiding them with deceptive perturbations. To enhance data unlearnability, we introduce a Latent Feature Concentration module, designed to work with the INN, regularizing the intra-class variance of these perturbations. To further boost the robustness of unlearnable examples, we design a Semantic Images Generation module that produces hidden semantic images. By utilizing similar semantic information, this module generates similar semantic images for samples within the same classes, thereby enlarging the inter-class distance and narrowing the intra-class distance. Extensive experiments on CIFAR-10, CIFAR-100, and an ImageNet subset, against 18 countermeasures, reveal that our proposed method exhibits outstanding robustness for unlearnable examples, demonstrating its efficacy in preventing unauthorized data exploitation. Nanyang Technological University This work was supported in part by the Rapid-Rich Object Search (ROSE) Laboratory, Nanyang Technological University in Singapore and the Peking University in China (NTU-PKU) Joint Research Institute (sponsored by the Ng Teng Fong Charitable Foundation), Nanyang Technological University, Singapore; and in part by the Defence Science Organisation of Singapore (DSO) National Laboratories under Project DSOCL22332. 2024-10-15T08:14:56Z 2024-10-15T08:14:56Z 2024 Journal Article Meng, R., Yi, C., Yu, Y., Yang, S., Shen, B. & Kot, A. C. (2024). Semantic deep hiding for robust unlearnable examples. IEEE Transactions On Information Forensics and Security, 19, 6545-6558. https://dx.doi.org/10.1109/TIFS.2024.3421273 1556-6013 https://hdl.handle.net/10356/180637 10.1109/TIFS.2024.3421273 2-s2.0-85197490122 19 6545 6558 en DSOCL22332 IEEE Transactions on Information Forensics and Security © 2024 IEEE. All rights reserved. |
institution |
Nanyang Technological University |
building |
NTU Library |
continent |
Asia |
country |
Singapore Singapore |
content_provider |
NTU Library |
collection |
DR-NTU |
language |
English |
topic |
Engineering Unlearnable examples Deep hiding |
spellingShingle |
Engineering Unlearnable examples Deep hiding Meng, Ruohan Yi, Chenyu Yu, Yi Yang, Siyuan Shen, Bingquan Kot, Alex C. Semantic deep hiding for robust unlearnable examples |
description |
Ensuring data privacy and protection has become paramount in the era of deep learning. Unlearnable examples are proposed to mislead the deep learning models and prevent data from unauthorized exploration by adding small perturbations to data. However, such perturbations (e.g., noise, texture, color change) predominantly impact low-level features, making them vulnerable to common countermeasures. In contrast, semantic images with intricate shapes have a wealth of high-level features, making them more resilient to countermeasures and potential for producing robust unlearnable examples. In this paper, we propose a Deep Hiding (DH) scheme that adaptively hides semantic images enriched with high-level features. We employ an Invertible Neural Network (INN) to invisibly integrate predefined images, inherently hiding them with deceptive perturbations. To enhance data unlearnability, we introduce a Latent Feature Concentration module, designed to work with the INN, regularizing the intra-class variance of these perturbations. To further boost the robustness of unlearnable examples, we design a Semantic Images Generation module that produces hidden semantic images. By utilizing similar semantic information, this module generates similar semantic images for samples within the same classes, thereby enlarging the inter-class distance and narrowing the intra-class distance. Extensive experiments on CIFAR-10, CIFAR-100, and an ImageNet subset, against 18 countermeasures, reveal that our proposed method exhibits outstanding robustness for unlearnable examples, demonstrating its efficacy in preventing unauthorized data exploitation. |
author2 |
School of Electrical and Electronic Engineering |
author_facet |
School of Electrical and Electronic Engineering Meng, Ruohan Yi, Chenyu Yu, Yi Yang, Siyuan Shen, Bingquan Kot, Alex C. |
format |
Article |
author |
Meng, Ruohan Yi, Chenyu Yu, Yi Yang, Siyuan Shen, Bingquan Kot, Alex C. |
author_sort |
Meng, Ruohan |
title |
Semantic deep hiding for robust unlearnable examples |
title_short |
Semantic deep hiding for robust unlearnable examples |
title_full |
Semantic deep hiding for robust unlearnable examples |
title_fullStr |
Semantic deep hiding for robust unlearnable examples |
title_full_unstemmed |
Semantic deep hiding for robust unlearnable examples |
title_sort |
semantic deep hiding for robust unlearnable examples |
publishDate |
2024 |
url |
https://hdl.handle.net/10356/180637 |
_version_ |
1814777808786817024 |