Towards secure and practical machine learning via secret sharing and random permutation

With the increasing demand for privacy protection, privacy-preserving machine learning has been drawing much attention from both academia and industry. However, most existing methods have their limitations in practical applications. On the one hand, although most cryptographic methods are provable secure, they bring heavy computation and communication. On the other hand, the security of many relatively efficient privacy-preserving techniques (e.g., federated learning and split learning) is being questioned, since they are non-provable secure. Inspired by previous work on privacy-preserving machine learning, we build a privacy-preserving machine learning framework by combining random permutation and arithmetic secret sharing via our compute-after-permutation technique. Our method is more efficient than existing cryptographic methods, since it can reduce the cost of element-wise function computation. Moreover, by adopting distance correlation as a metric for evaluating privacy leakage, we demonstrate that our method is more secure than previous non-provable secure methods. Overall, our proposal achieves a good balance between security and efficiency. Experimental results show that our method not only is up to 5× faster and reduces up to 80% network traffic compared with state-of-the-art cryptographic methods, but also leaks less privacy during the training process compared with non-provable secure methods.