19 August 2024 in IEEE Transactions on Dependable and Secure Computing
Considerable efforts have been devoted to addressing distributed backdoor attacks in federated learning (FL) systems. While significant progress has been made in enhancing the security of FL systems, our study reveals that there remains a false sense of security surrounding FL. We demonstrate that colluding malicious participants can effectively execute backdoor attacks during the FL training process, exhibiting high sparsity and stealthiness, which means they can evade common defense methods with only a few attack iterations. Our research highlights this vulnerability by proposing a Co llusive B ackdoor A ttack named CoBA . CoBA is designed to enhance the sparsity and stealthiness of backdoor attacks by offering trigger tuning to facilitate learning of backdoor training data, controlling the bias of malicious local model updates, and applying the projected gradient descent technique. By conducting extensive empirical studies on 5 benchmark datasets, we make the following observations: 1) CoBA successfully circumvents 15 state-of-the-art defense methods for robust FL; 2) Compared to existing backdoor attacks, CoBA consistently achieves superior attack performance; and 3) CoBA can achieve persistent poisoning effects through significantly sparse attack iterations. These findings raise substantial concerns regarding the integrity of FL and underscore the urgent need for heightened vigilance in defending against such attacks.
1 June 2024 in ACM Transactions on Sensor Networks
Federated learning (FL), which holds promise for use in edge intelligence applications for smart cities, enables smart devices collaborate in training a global model by exchanging local model updates instead of sharing local training data. However, the global model can be corrupted by malicious clients conducting poisoning attacks, resulting in the failure of converging the global model, incorrect predictions on the test set, or the backdoor embedded. Although some aggregation algorithms can enhance the robustness of FL against malicious clients, our work demonstrates that existing stealthy poisoning attacks can still bypass these defense methods. In this work, we propose a robust aggregation mechanism, called Parts and All ( PnA ), to protect the global model of FL by filtering out malicious local model updates throughout the detection of poisoning attacks at layers of local model updates. We conduct comprehensive experiments on three representative datasets. The experimental results demonstrate that our proposed PnA is more effective than existing robust aggregation algorithms against state-of-the-art poisoning attacks. Besides, PnA has a stable performance against poisoning attacks with different poisoning settings.
10 June 2026 in IEEE Transactions on Dependable and Secure Computing
Graph Prompt Learning (GPL) effectively bridges the gap between graph pre-training and downstream tasks, addressing the bottleneck in knowledge transfer. However, its susceptibility to backdoor attacks during pre-training remains largely unexplored. This work systematically investigates the backdoor risks in GPL and introduces TrojanGPL-a cross-context backdoor attack that embeds triggers solely through manipulations of the pre-trained encoder, independent of downstream knowledge. Our analysis reveals that jointly optimizing trigger graphs with the prompt mechanism not only narrows the gap between upstream and downstream main tasks but also aligns main and backdoor learning objectives, facilitating seamless backdoor transfer from pre-trained models to diverse downstream applications. To counter this threat, we propose ES-Detector, a defense mechanism that identifies backdoored graph samples by detecting discriminative embedding shifts upon the removal of important nodes. Specifically, when clean and backdoored graphs are predicted as the same class, clean graphs exhibit stable embedding changes, whereas backdoored graphs show significant deviations from clean ones. Extensive experiments on 5 benchmark datasets and 3 GPL frameworks demonstrate that TrojanGPL achieves an attack success rate exceeding 90% in different cross-context scenarios while minimally impacting the performance of the main task. Meanwhile, ES-Detector reduces the attack success rate by at least 60% across all the test settings, validating its practical robustness without prior knowledge of attack patterns.