Lightweight privacy protection generative adversarial network system

Abstract

The invention relates to a lightweight privacy protection generative adversarial network system; an entity comprises a data provider DS, a service provider SP, a first edge server S1 and a second edgeserver S2, and software comprises an LP-GAN security computing framework; the LP-GAN security computing framework comprises a security generation model SG and a security discrimination model SD. According to the invention, dual privacy of data and models of users can be ensured.

Description

technical field [0001] The invention relates to the field of machine learning, in particular to a lightweight privacy protection generating confrontation network system. Background technique [0002] Deep learning has achieved important breakthroughs in a variety of computer vision tasks, such as image recognition, image translation, and medical image diagnosis. In order to improve the accuracy of the model, deep learning algorithms usually require a large number of training samples. However, due to privacy and legal restrictions, there is not enough training data in many data-sensitive fields (such as medical and military fields), and missing data makes it more difficult to build models. [0003] Generative confrontation network (GAN) has the ability to learn the distribution of real data (such as images, texts, videos), and can generate generated samples that are difficult to distinguish from real samples, especially in generating synthetic images than many traditional fr...

AI Technical Summary

Problems solved by technology

However, the existing framework based on homomorphic encryption has high time complexity and large storage consumption, which is not suitable for practical applications

Images