A data protection method based on gan-based federated learning intrusion detection

Abstract

The invention belongs to the technical field of intrusion detection, and in particular relates to a data protection method for intrusion detection based on GAN federated learning. The invention combines the loose differential privacy protection technology with the adversarial generation neural network, reduces the communication loss of each terminal in the federated learning framework, improves the learning efficiency, can well solve the situation that the computing power of each federated terminal is low, and improves the utilization of the machine efficiency. The dynamic fuzzy data generated by the adversarial generation neural network used in the present invention can prevent the attacker from judging the success of the attack while expanding the local training data set to solve the possible small sample problem. The present invention can effectively reduce the communication loss in the federated learning framework, can effectively improve the training efficiency, and at the same time solve the problem of being vulnerable to reasoning attacks in the federated learning of few terminals, the small sample data of intrusion detection terminals and the problem of non-independent and identically distributed data , which can implement federated learning intrusion detection against inference attacks.

Description

technical field [0001] The invention belongs to the technical field of intrusion detection, and in particular relates to a data protection method for intrusion detection based on GAN federated learning. Background technique [0002] Federated learning is a safe distributed learning mode. It realizes the shared model by continuously interacting model parameters between the client and the parameter server, so as to realize the sharing of data resources without leaving the client and improve the efficiency of data utilization. At the same time, the data privacy of users is guaranteed. [0003] Federated learning was created to solve user privacy issues. It can realize privacy-protected joint machine learning without affecting efficiency as much as possible. Therefore, the security of federated learning itself is very important. The current security problems of federated learning come from two parts, one is the data security problem in the process of data transmission, and the ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to overcome the problem that the current federated learning intrusion detection communication overhead is high, and the use of differential privacy protection cannot effectively defend against reasoning attacks when there are fewer clients, and provide a data protection based on GAN federated learning intrusion detection method

Images