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

Deep Learning-based image synthesis techniques have been applied inhealthcare research for generating medical images to support open research andaugment medical datasets. Training generative adversarial neural networks(GANs) usually require large amounts of training data. Federated learning (FL)provides a way of training a central model using distributed data while keepingraw data locally. However, given that the FL server cannot access the raw data,it is vulnerable to backdoor attacks, an adversarial by poisoning trainingdata. Most backdoor attack strategies focus on classification models andcentralized domains. It is still an open question if the existing backdoorattacks can affect GAN training and, if so, how to defend against the attack inthe FL setting. In this work, we investigate the overlooked issue of backdoorattacks in federated GANs (FedGANs). The success of this attack is subsequentlydetermined to be the result of some local discriminators overfitting thepoisoned data and corrupting the local GAN equilibrium, which then furthercontaminates other clients when averaging the generator's parameters and yieldshigh generator loss. Therefore, we proposed FedDetect, an efficient andeffective way of defending against the backdoor attack in the FL setting, whichallows the server to detect the client's adversarial behavior based on theirlosses and block the malicious clients. Our extensive experiments on twomedical datasets with different modalities demonstrate the backdoor attack onFedGANs can result in synthetic images with low fidelity. After detecting andsuppressing the detected malicious clients using the proposed defense strategy,we show that FedGANs can synthesize high-quality medical datasets (with labels)for data augmentation to improve classification models' performance.

Backdoor Attack and Defense in Federated Generative Adversarial Network-based Medical Image Synthesis