Phillip Rieger (Technical University of Darmstadt), Thien Duc Nguyen (Technical University of Darmstadt), Markus Miettinen (Technical University of Darmstadt), Ahmad-Reza Sadeghi (Technical University of Darmstadt)

Federated Learning (FL) allows multiple clients to collaboratively train a Neural Network (NN) model on their private data without revealing the data. Recently, several targeted poisoning attacks against FL have been introduced. These attacks inject a backdoor into the resulting model that allows adversary-controlled inputs to be misclassified. Existing countermeasures against backdoor attacks are inefficient and often merely aim to exclude deviating models from the aggregation. However, this approach also removes benign models of clients with deviating data distributions, causing the aggregated model to perform poorly for such clients.

To address this problem, we propose emph{DeepSight}, a novel model filtering approach for mitigating backdoor attacks. It is based on three novel techniques that allow to characterize the distribution of data used to train model updates and seek to measure fine-grained differences in the internal structure and outputs of NNs. Using these techniques,
DeepSight can identify suspicious model updates. We also develop a scheme that can accurately cluster model updates. Combining the results of both components, DeepSight is able to identify and eliminate model clusters containing poisoned models with high attack impact. We also show that the backdoor contributions of possibly undetected poisoned models can be effectively mitigated with existing weight clipping-based defenses. We evaluate the performance and effectiveness of DeepSight and show that it can mitigate state-of-the-art backdoor attacks with a negligible impact on the model's performance on benign data.

View More Papers

Shaduf: Non-Cycle Payment Channel Rebalancing

Zhonghui Ge (Shanghai Jiao Tong University), Yi Zhang (Shanghai Jiao Tong University), Yu Long (Shanghai Jiao Tong University), Dawu Gu (Shanghai Jiao Tong University)

Read More

Clarion: Anonymous Communication from Multiparty Shuffling Protocols

Saba Eskandarian (University of North Carolina at Chapel Hill), Dan Boneh (Stanford University)

Read More

Demo #2: Policy-based Discovery and Patching of Logic Bugs...

Hyungsub Kim (Purdue University), Muslum Ozgur Ozmen (Purdue University), Antonio Bianchi (Purdue University), Z. Berkay Celik (Purdue University) and Dongyan Xu (Purdue University)

Read More

Hazard Integrated: Understanding Security Risks in App Extensions to...

Mingming Zha (Indiana University Bloomington), Jice Wang (National Computer Network Intrusion Protection Center, University of Chinese Academy of Sciences), Yuhong Nan (Sun Yat-sen University), Xiaofeng Wang (Indiana Unversity Bloomington), Yuqing Zhang (National Computer Network Intrusion Protection Center, University of Chinese Academy of Sciences), Zelin Yang (National Computer Network Intrusion Protection Center, University of Chinese Academy…

Read More