A method of adversarial sample generation based on geometric vector

Abstract

The invention discloses a method for generating an adversarial sample based on a geometric vector, comprising: step 1, data preprocessing; step 2, model pre-training; step 3, repeating steps (a)-(f) until convergence, to obtain the DGA domain name Adversarial example: (a) input the legal domain name into the ATN network to generate the legal domain name adversarial sample, and obtain the disturbance loss; (b) input the legal domain name and the legal domain name adversarial sample into the noise disturbance direction function to obtain the noise; (c) combine the noise and the DGA The domain name is input into the disturbance network to obtain the DGA domain name confrontation sample; the disturbance network is a disturbance network based on a geometric vector; (d) inputting the DGA domain name confrontation sample into the target network to obtain the target network loss; (e) using the disturbance loss and the target network loss to obtain Target loss function; (f) Update the ATN network by minimizing the target loss function. The present invention can generate its adversarial examples for a specific DGA category.

Description

An Adversarial Sample Generation Method Based on Geometric Vectors technical field [0001] The present invention relates to the technical field of machine learning, in particular to a method for generating adversarial samples based on geometric vectors. Background technique [0002] Deep neural networks perform very well in complex tasks, but recent studies have shown that they can Vulnerable to adversarial attacks, a form of attack that adds tiny perturbations to the input that cause the model to predict incorrectly output. In practical applications, adversarial attacks pose a serious threat to the success of deep learning. For this reason, the researchers put forward A method to generate samples to deal with potential attacks and enhance the robustness and generalization ability of neural networks. [0003] At present, in the principle of adversarial sample generation, it is mainly divided into two categories, one is randomly added to the original sample Noise until...

AI Technical Summary

Problems solved by technology

However, these methods have the following problems: not any perturbation of the original sample will cause classification errors, only specific perturbations will cause classification errors, so by adding noise randomly, it is not easy to get adversarial samples

Although the gradient information gives the perturbation direction, for high-dimensional data and more complex neural network models, it will increase the complexity of establishing gradient information

Existing adversarial sample generation methods do not take into account the cost and efficiency of generating samples

Images