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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

This work explores an emerging security threat against deep neural networks(DNNs) based image classification, i.e., backdoor attack. In this scenario, theattacker aims to inject a backdoor into the model by manipulating trainingdata, such that the backdoor could be activated by a particular trigger andbootstraps the model to make a target prediction at inference. Currently, mostexisting data poisoning-based attacks struggle to achieve success at lowpoisoning ratios, increasing the risk of being defended by defense methods. Inthis paper, we propose a novel frequency-based backdoor attack via WaveletPacket Decomposition (WPD), WPD decomposes the original image signal to aspectrogram that contains frequency information with different semanticmeanings. We leverage WPD to statistically analyze the frequency distributionof the dataset to infer the key frequency regions the DNNs would focus on, andthe trigger information is only injected into the key frequency regions. Ourmethod mainly includes three parts: 1) the selection of the poisoning frequencyregions in spectrogram; 2) trigger generation; 3) the generation of thepoisoned dataset. Our method is stealthy and precise, evidenced by the 98.12%Attack Success Rate (ASR) on CIFAR-10 with the extremely low poisoning ratio0.004% (i.e., only 2 poisoned samples among 50,000 training samples) and canbypass most existing defense methods. Besides, we also provide visualizationanalyses to explain why our method works.

WPDA: Frequency-based Backdoor Attack with Wavelet Packet Decomposition