About: Data-Free Adversarial Perturbations for Practical Black-Box Attack

Facets (new session)
Description
Metadata
Settings
- owl:sameAs
- Inference Rule:

About: Data-Free Adversarial Perturbations for Practical Black-Box Attack Goto Sponge NotDistinct Permalink

An Entity of Type : schema:ScholarlyArticle, within Data Space : wasabi.inria.fr associated with source document(s)

Attributes	Values
type	Academic Article research paper schema:ScholarlyArticle
isDefinedBy	Covid-on-the-Web dataset
has title	Data-Free Adversarial Perturbations for Practical Black-Box Attack
Creator	Zhou, Jun Zhang, Xiaolu Fu, Chilin Huan, Zhaoxin Shang, Lin Wang, Yulong
Source	PMC
abstract	Neural networks are vulnerable to adversarial examples, which are malicious inputs crafted to fool pre-trained models. Adversarial examples often exhibit black-box attacking transferability, which allows that adversarial examples crafted for one model can fool another model. However, existing black-box attack methods require samples from the training data distribution to improve the transferability of adversarial examples across different models. Because of the data dependence, fooling ability of adversarial perturbations is only applicable when training data are accessible. In this paper, we present a data-free method for crafting adversarial perturbations that can fool a target model without any knowledge about the training data distribution. In the practical setting of black-box attack scenario where attackers do not have access to target models and training data, our method achieves high fooling rates on target models and outperforms other universal adversarial perturbation methods. Our method empirically shows that current deep learning models are still at a risk even when the attackers do not have access to training data.
has issue date	2020-04-17(xsd:dateTime)
bibo:doi	10.1007/978-3-030-47436-2_10
has license	no-cc
sha1sum (hex)	ae16ad10188b09e5fb06c166a2f0121d92d597f9
schema:url	https://doi.org/10.1007/978-3-030-47436-2_10
resource representing a document's title	Data-Free Adversarial Perturbations for Practical Black-Box Attack
has PubMed Central identifier	PMC7206253
schema:publication	Advances in Knowledge Discovery and Data Mining
resource representing a document's body	covid:ae16ad10188b09e5fb06c166a2f0121d92d597f9#body_text
is schema:about of	named entity 'access' covid:arg/ae16ad10188b09e5fb06c166a2f0121d92d597f9 named entity 'models' named entity 'However' named entity 'perturbation methods' named entity 'models' named entity 'training data' named entity 'perturbation methods' named entity 'training data' named entity 'ImageNet' named entity 'training data' named entity 'softmax' named entity 'machine learning' named entity 'transfer-based' named entity 'training data' named entity '60,000' named entity 'API2' named entity 'pre-trained' named entity 'training data' named entity 'neurons' named entity 'probability' named entity 'pre-trained' named entity 'training data' named entity '1, 2' named entity 'Algorithm' named entity 'FGSM' named entity 'objective function' named entity 'softmax function' named entity 'decision boundaries' named entity 'training data' named entity '1, 2' named entity 'features extracted' named entity 'ImageNet' named entity 'representation space' named entity 'objective function' named entity 'loss function' named entity '3, 4' named entity 'classifiers' named entity 'pre-trained' named entity 'training data' named entity 'deep learning' named entity 'pre-trained' named entity 'Kullback-Leibler' named entity 'pre-trained' named entity 'agnostic' named entity 'pre-trained' named entity 'pre-trained' named entity 'objective function' named entity 'softmax' named entity 'medical record' named entity 'training data' named entity 'CIFAR-10' named entity 'FGSM' named entity 'pre-trained' named entity 'training data' named entity 'probability' named entity 'ResNet' named entity 'deep neural networks' named entity 'numerical optimization' named entity 'data sample' named entity 'ResNet' named entity 'training data' named entity 'objective functions' named entity 'sampling' named entity '3, 4' named entity 'pre-trained' named entity 'agnostic' named entity 'back-propagation' named entity 'pre-trained' named entity 'training set' named entity 'GoogleNet' named entity 'academic research' named entity 'ImageNet' named entity 'ImageNet' named entity 'training data' named entity 'CIFAR-10' named entity 'logit' named entity 'neuron'

Faceted Search & Find service v1.13.91 as of Mar 24 2020

Alternative Linked Data Documents: Sponger | ODE Content Formats:

RDF

ODATA

Microdata

About

OpenLink Virtuoso version 07.20.3229 as of Jul 10 2020, on Linux (x86_64-pc-linux-gnu), Single-Server Edition (94 GB total memory)
Data on this page belongs to its respective rights holders.
Virtuoso Faceted Browser Copyright © 2009-2024 OpenLink Software