Deep neural networks (DNNs) are phenomenally successful machine learning methods broadly applied to many different disciplines. However, as complex two-party computations, DNN inference using classical cryptographic methods cannot achieve unconditional security, raising concern on security risks of DNNs' ...
In this section, we introduce our proposed secure neural network inference network that combines the B-LNN with A-SS. We then report the evaluation results of our framework and compare them with previous works. Conclusion In this work, we investigate the high latency problem bought by activation...
客户持有输入x, 希望得到f(x), 但是客户希望x是保密的, 而模型提供商也不希望与客户共享NN模型f的信息. 安全神经网络推理(secure neural networks inference, SNNI)问题即是在满足这些安全需求的同时计算f(x). 有些文献也把安全推理称为隐私保护推理(privacy-preserving inference), 不经意预测(oblivious...
今天给大家带来的是来自于USENIX Security'22的一篇文章《Cheetah: Lean and Fast Secure Two-Party Deep Neural Network Inference》, 作者是来自阿里安全双子座实验室的洪澄博士团队, 文章链接如下: Cheetah: Lean and Fast Secure Two-Party Deep Neural Network Inferenceeprint.iacr.org/2022/207 摘要 2PC-...
内容提示: GAZELLE : A Low Latency Framework for SecureNeural Network InferenceChiraag JuvekarMIT MTLchiraag@mit.eduVinod VaikuntanathanMIT CSAILvinodv@csail.mit.eduAnantha ChandrakasanMIT MTLanantha@mtl.mit.eduAbstract—The growing popularity of cloud-based machinelearning raises a natural question about ...
To this end, we design Gazelle, a scalable and low-latency system for secure neural network inference, using an intricate combination of homomorphic encryption and traditional two-party computation techniques (such as garbled circuits). Gazelle makes three contributions. First, we design the Gazelle ...
while guaranteeing the privacy of the server's neural network. To this end, we design Gazelle, a scalable and low-latency system for secure neural network inference, using an intricate combination of homomorphic encryption and traditional two-party computation techniques (such as garbled circuits). ...
[10] used LHE to design a secure neural network inference protocol. However, the scheme suffered from the high computational cost [11]. The recent work in [11] utilized a Packed Additive Homomorphic Encryption (PAHE) scheme to directly speed up linear algebra which has shown its effectiveness....
(2018). \(\{\)GAZELLE\(\}\): A low latency framework for secure neural network inference. In: 27th \(\{\)USENIX\(\}\) Security Symposium (\(\{\)USENIX\(\}\) Security 18), pp. 1651–1669 Kairouz, P., Oh, S., & Viswanath, P. (2016). Extremal mechanisms for local ...
Juvekar C, Vaikuntanathan V, Chandrakasan A (2018) GAZELLE: a low latency framework for secure neural network inference. In: 27th USENIX security symposium (USENIX Security 18). USENIX Association, Baltimore, pp 1651–1669 Khayyam H, Javadi B, Jalili M, Jazar RN (2020) Artificial intelligence...