Janus: Dual-Server Multi-Round Secure Aggregation with Verifiability for Federated Learning

Lang Pu, Jingjing Gu, Chao Lin, Xinyi Huang
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:49927-49945, 2025.

Abstract

Secure Aggregation (SA) is a cornerstone of Federated Learning (FL), ensuring that user updates remain hidden from servers. The advanced Flamingo (S&P’23) has realized multi-round aggregation and improved efficiency. However, it still faces several key challenges: scalability issues with dynamic user participation, a lack of verifiability for server-side aggregation results, and vulnerability to Model Inconsistency Attacks (MIA) caused by a malicious server distributing inconsistent models. To address these issues, we propose $\textit{Janus}$, a generic SA scheme based on dual-server architecture. Janus ensures security against up to $n-2$ colluding clients (where $n$ is the total client count), which prevents privacy breaches for non-colluders. Additionally, Janus is model-independent, ensuring applicability across any FL model without specific adaptations. Furthermore, Janus introduces a new cryptographic primitive, Separable Homomorphic Commitment, which enables clients to efficiently verify the correctness of aggregation. Finally, extensive experiments show that Janus not only significantly enhances security but also reduces per-client communication and computation overhead from logarithmic to constant scale, with a tolerable impact on model performance.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-pu25a, title = {Janus: Dual-Server Multi-Round Secure Aggregation with Verifiability for Federated Learning}, author = {Pu, Lang and Gu, Jingjing and Lin, Chao and Huang, Xinyi}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {49927--49945}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/pu25a/pu25a.pdf}, url = {https://proceedings.mlr.press/v267/pu25a.html}, abstract = {Secure Aggregation (SA) is a cornerstone of Federated Learning (FL), ensuring that user updates remain hidden from servers. The advanced Flamingo (S&P’23) has realized multi-round aggregation and improved efficiency. However, it still faces several key challenges: scalability issues with dynamic user participation, a lack of verifiability for server-side aggregation results, and vulnerability to Model Inconsistency Attacks (MIA) caused by a malicious server distributing inconsistent models. To address these issues, we propose $\textit{Janus}$, a generic SA scheme based on dual-server architecture. Janus ensures security against up to $n-2$ colluding clients (where $n$ is the total client count), which prevents privacy breaches for non-colluders. Additionally, Janus is model-independent, ensuring applicability across any FL model without specific adaptations. Furthermore, Janus introduces a new cryptographic primitive, Separable Homomorphic Commitment, which enables clients to efficiently verify the correctness of aggregation. Finally, extensive experiments show that Janus not only significantly enhances security but also reduces per-client communication and computation overhead from logarithmic to constant scale, with a tolerable impact on model performance.} }
Endnote
%0 Conference Paper %T Janus: Dual-Server Multi-Round Secure Aggregation with Verifiability for Federated Learning %A Lang Pu %A Jingjing Gu %A Chao Lin %A Xinyi Huang %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-pu25a %I PMLR %P 49927--49945 %U https://proceedings.mlr.press/v267/pu25a.html %V 267 %X Secure Aggregation (SA) is a cornerstone of Federated Learning (FL), ensuring that user updates remain hidden from servers. The advanced Flamingo (S&P’23) has realized multi-round aggregation and improved efficiency. However, it still faces several key challenges: scalability issues with dynamic user participation, a lack of verifiability for server-side aggregation results, and vulnerability to Model Inconsistency Attacks (MIA) caused by a malicious server distributing inconsistent models. To address these issues, we propose $\textit{Janus}$, a generic SA scheme based on dual-server architecture. Janus ensures security against up to $n-2$ colluding clients (where $n$ is the total client count), which prevents privacy breaches for non-colluders. Additionally, Janus is model-independent, ensuring applicability across any FL model without specific adaptations. Furthermore, Janus introduces a new cryptographic primitive, Separable Homomorphic Commitment, which enables clients to efficiently verify the correctness of aggregation. Finally, extensive experiments show that Janus not only significantly enhances security but also reduces per-client communication and computation overhead from logarithmic to constant scale, with a tolerable impact on model performance.
APA
Pu, L., Gu, J., Lin, C. & Huang, X.. (2025). Janus: Dual-Server Multi-Round Secure Aggregation with Verifiability for Federated Learning. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:49927-49945 Available from https://proceedings.mlr.press/v267/pu25a.html.

Related Material