ScionFL: Efficient and Robust Secure Quantized Aggregation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2022
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910450054070272 |
|---|---|
| author | Ben-Itzhak, Yaniv Möllering, Helen Pinkas, Benny Schneider, Thomas Suresh, Ajith Tkachenko, Oleksandr Vargaftik, Shay Weinert, Christian Yalame, Hossein Yanai, Avishay |
| author_facet | Ben-Itzhak, Yaniv Möllering, Helen Pinkas, Benny Schneider, Thomas Suresh, Ajith Tkachenko, Oleksandr Vargaftik, Shay Weinert, Christian Yalame, Hossein Yanai, Avishay |
| contents | Secure aggregation is commonly used in federated learning (FL) to alleviate privacy concerns related to the central aggregator seeing all parameter updates in the clear. Unfortunately, most existing secure aggregation schemes ignore two critical orthogonal research directions that aim to (i) significantly reduce client-server communication and (ii) mitigate the impact of malicious clients. However, both of these additional properties are essential to facilitate cross-device FL with thousands or even millions of (mobile) participants.
In this paper, we unite both research directions by introducing ScionFL, the first secure aggregation framework for FL that operates efficiently on quantized inputs and simultaneously provides robustness against malicious clients. Our framework leverages (novel) multi-party computation (MPC) techniques and supports multiple linear (1-bit) quantization schemes, including ones that utilize the randomized Hadamard transform and Kashin's representation.
Our theoretical results are supported by extensive evaluations. We show that with no overhead for clients and moderate overhead for the server compared to transferring and processing quantized updates in plaintext, we obtain comparable accuracy for standard FL benchmarks. Moreover, we demonstrate the robustness of our framework against state-of-the-art poisoning attacks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2210_07376 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | ScionFL: Efficient and Robust Secure Quantized Aggregation Ben-Itzhak, Yaniv Möllering, Helen Pinkas, Benny Schneider, Thomas Suresh, Ajith Tkachenko, Oleksandr Vargaftik, Shay Weinert, Christian Yalame, Hossein Yanai, Avishay Cryptography and Security Information Theory Machine Learning Secure aggregation is commonly used in federated learning (FL) to alleviate privacy concerns related to the central aggregator seeing all parameter updates in the clear. Unfortunately, most existing secure aggregation schemes ignore two critical orthogonal research directions that aim to (i) significantly reduce client-server communication and (ii) mitigate the impact of malicious clients. However, both of these additional properties are essential to facilitate cross-device FL with thousands or even millions of (mobile) participants. In this paper, we unite both research directions by introducing ScionFL, the first secure aggregation framework for FL that operates efficiently on quantized inputs and simultaneously provides robustness against malicious clients. Our framework leverages (novel) multi-party computation (MPC) techniques and supports multiple linear (1-bit) quantization schemes, including ones that utilize the randomized Hadamard transform and Kashin's representation. Our theoretical results are supported by extensive evaluations. We show that with no overhead for clients and moderate overhead for the server compared to transferring and processing quantized updates in plaintext, we obtain comparable accuracy for standard FL benchmarks. Moreover, we demonstrate the robustness of our framework against state-of-the-art poisoning attacks. |
| title | ScionFL: Efficient and Robust Secure Quantized Aggregation |
| topic | Cryptography and Security Information Theory Machine Learning |
| url | https://arxiv.org/abs/2210.07376 |