Highly Accurate CNN Inference Using Approximate Activation Functions over Homomorphic Encryption
In the big data era, cloud-based machine learning as a service (MLaaS) has attracted considerable attention. However, when handling sensitive data, such as financial and medical data, a privacy issue emerges, because the cloud server can access clients' raw data. A common method of handling sensitive data in the cloud uses homomorphic encryption, which allows computation over encrypted data without decryption. Previous research usually adopted a low-degree polynomial mapping function, such as the square function, for data classification. However, this technique results in low classification accuracy. In this study, we seek to improve the classification accuracy for inference processing in a convolutional neural network (CNN) while using homomorphic encryption. We adopt an activation function that approximates Google's Swish activation function while using a fourth-order polynomial. We also adopt batch normalization to normalize the inputs for the Swish function to fit the input range to minimize the error. We implemented CNN inference labeling over homomorphic encryption using the Microsoft's Simple Encrypted Arithmetic Library for the Cheon-Kim-Kim-Song (CKKS) scheme. The experimental evaluations confirmed classification accuracies of 99.22% and 80.48% for MNIST and CIFAR-10, respectively, which entails 0.04% and 4.11% improvements, respectively, over previous methods.
Code (0)
등록된 구현이 없습니다.
Tasks
ClassificationGeneral ClassificationMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
A Constant-Time Implementation Methodology for Activation Functions on Microcontrollers
Embedded neural-network inference can leak information through timing side channels, including leakage caused by the evaluation of activation functions. This work proposes a constant-time implementation methodology for a…
A More Accurate Approximation of Activation Function with Few Spikes Neurons
Recent deep neural networks (DNNs), such as diffusion models [1], have faced high computational demands. Thus, spiking neural networks (SNNs) have attracted lots of attention as energy-efficient neural networks. However,…
K-TanH: Efficient TanH For Deep Learning
We propose K-TanH, a novel, highly accurate, hardware efficient approximation of popular activation function TanH for Deep Learning. K-TanH consists of parameterized low-precision integer operations, such as, shift and a…
CPUDeep LearningTranslationCombU: A Combined Unit Activation for Fitting Mathematical Expressions with Neural Networks
The activation functions are fundamental to neural networks as they introduce non-linearity into data relationships, thereby enabling deep networks to approximate complex data relations. Existing efforts to enhance neura…
Neural Networks with (Low-Precision) Polynomial Approximations: New Insights and Techniques for Accuracy Improvement
Replacing non-polynomial functions (e.g., non-linear activation functions such as ReLU) in a neural network with their polynomial approximations is a standard practice in privacy-preserving machine learning. The resultin…
Privacy Preserving