English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

In this paper, we present Zero-data Based Repeated bit flip Attack (ZeBRA)that precisely destroys deep neural networks (DNNs) by synthesizing its ownattack datasets. Many prior works on adversarial weight attack require not onlythe weight parameters, but also the training or test dataset in searchingvulnerable bits to be attacked. We propose to synthesize the attack dataset,named distilled target data, by utilizing the statistics of batch normalizationlayers in the victim DNN model. Equipped with the distilled target data, ourZeBRA algorithm can search vulnerable bits in the model without accessingtraining or test dataset. Thus, our approach makes the adversarial weightattack more fatal to the security of DNNs. Our experimental results show that2.0x (CIFAR-10) and 1.6x (ImageNet) less number of bit flips are required onaverage to destroy DNNs compared to the previous attack method. Our code isavailable at https://github. com/pdh930105/ZeBRA.

ZeBRA: Precisely Destroying Neural Networks with Zero-Data Based Repeated Bit Flip Attack