Confrontation sample generation method and device, storage medium, equipment and program product

Abstract

The invention discloses an adversarial sample generation method and device, a storage medium, equipment and a program product, and can be applied to scenes such as artificial intelligence, machine learning and model attack adversarial. The method comprises the following steps: training a to-be-trained model by adopting an open source data set to obtain a pre-trained model; the target data set is adopted to adjust the pre-training model to obtain a proxy model, the target data set comprises a target sample and a target classification label, the target classification label is obtained by classifying the target sample by the target model, and the proxy model can simulate the decision behavior of the target model; and generating an adversarial sample according to the proxy model and a preset white-box attack algorithm, wherein the adversarial sample can be used for attacking the target model. According to the embodiment of the invention, the open source data set is adopted to pre-train the to-be-trained model, so that the sample size of the target data set required for obtaining the proxy model can be reduced, the target model does not need to be accessed for multiple times to obtain the target data set, and an attack behavior is not easy to discover.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, and in particular, to a method for generating an adversarial sample, an apparatus for generating an adversarial sample, a computer-readable storage medium, a computer device, and a computer program product. Background technique [0002] In order to test and evaluate product security, adversarial samples can be used to attack the target model of the product to determine whether the target model is easily affected by adversarial samples. In the related art, in order to generate adversarial samples, it is necessary to access the target model multiple times, which is easily identified as abnormal access traffic, resulting in the discovery of attack behaviors. SUMMARY OF THE INVENTION [0003] The embodiments of the present application provide a method for generating an adversarial sample, an apparatus for generating an adversarial sample, a computer-readable storage medium...

AI Technical Summary

Problems solved by technology

In related technologies, in order to generate adversarial samples, the target model needs to be accessed multiple times, so it is easy to be identified as abnormal access traffic, causing the attack behavior to be discovered

Images