Neural network model training method and device and secure computing platform

Abstract

The embodiment of the invention provides a neural network model training method and device and a device for neural network model training. The method comprises the following steps: acquiring training data from a participant, wherein the training data is a ciphertext; initializing a neural network model by executing the training script, wherein a network layer of the neural network model comprises a plaintext layer and a ciphertext layer; and carrying out iterative training on the neural network model based on the held training data, in each round of iterative training, carrying out security calculation on the training data of the round layer by layer through each network layer of the neural network model, and when an iteration termination condition is not met, according to a model output result, adjusting parameters of each network layer layer by layer through back propagation, so as to obtain the security of the neural network model. When an iteration termination condition is reached, obtaining a trained neural network model; calculation of the plaintext layer is carried out based on plaintext, and calculation of the ciphertext layer is carried out based on ciphertext. According to the embodiment of the invention, on the basis of ensuring data privacy security, communication overhead and resource consumption caused by ciphertext calculation are reduced.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to a neural network model training method, device and secure computing platform. Background technique [0002] With the wide application of machine learning, methods based on machine learning or deep learning have been applied in many technical fields and have achieved great success. When training each machine learning model, the data required for machine learning often involves multiple fields. [0003] However, due to factors such as information security and user privacy protection, different organizations or individuals hold different types of data, which are usually isolated from each other, making the data unable to circulate efficiently and conveniently, forming information islands. Therefore, under the premise of taking into account data security and computational efficiency, how to combine multiple data sources to jointly train a machine learning model is an urgen...

AI Technical Summary

Problems solved by technology

[0003] However, due to factors such as information security and user privacy protection, different organizations or individuals have different types of data, which are usually isolated from each other, making the data unable to circulate efficiently and conveniently, forming information islands

Images