Efficient Evaluation of Activation Functions over Encrypted Data
2019 IEEE Security and Privacy Workshops (SPW)(2019)
摘要
We describe a method for approximating any bounded activation function given encrypted input data. The utility of our method is exemplified by simulating it within two typical machine learning tasks: namely, a Variational Autoencoder that learns a latent representation of MNIST data, and an MNIST image classifier.
更多查看译文
关键词
machine learning,homomorphic encryption,security,privacy,non-polynomial function,activation function
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络