Decentralized federated machine learning method under privacy protection

Abstract

The invention discloses a decentralized federated learning method under privacy protection. The decentralized federated learning method comprises a system initialization step, a request model and local parallel training step, a model parameter encryption and model sending step, a model receiving and recovering step and a system updating step. Decentralization is achieved by using a strategy of randomly selecting participants as parameter aggregators, and the defects that existing federated learning is easily attacked by DoS, a parameter server has a single point of failure and the like are overcome; a PVSS verifiable secret distribution protocol is combined to protect participant model parameters from model inversion attacks and data member reasoning attacks. Meanwhile, it is guaranteed that parameter aggregation is carried out by different participants in each training task, when an untrusted aggregator occurs or the aggregator is attacked, the aggregator can recover to be normal by itself, and the robustness of federated learning is improved; while the functions are achieved, the federated learning performance is guaranteed, the safety training environment of federated learning is effectively improved, and wide application prospects are achieved.

Description

technical field [0001] The invention belongs to the intersection field of machine learning and information security, and in particular relates to a centralized federated learning method under privacy protection. Background technique [0002] Machine learning technology has made extraordinary achievements in artificial intelligence application scenarios such as face recognition, speech recognition, and natural language processing. However, how to ensure that machine learning operates in a safe environment is still an unresolved problem. The essence of machine learning is to use a large amount of data to train the algorithm model, and obtain an algorithm model (hereinafter referred to as the model) that can accurately predict new input data. The data sets used for machine learning contain a large amount of private data of users, such as personal pictures, medical insurance records, input method records, and so on. However, machine learning requires powerful computing support....

AI Technical Summary

Problems solved by technology

But if there is a backdoor program, then the differential privacy technology will lose its protective effect

In addition, when some auxiliary information is obtained by the attacker, it will also lead to data leakage

More importantly, there is a close relationship between data in real life. In this case, just setting the granularity of differential privacy cannot effectively protect privacy

[0006] In addition to the above-mentioned problems, many implementation forms of federated learning at this stage do not really take into account the attack behavior among users. This mutual distrust will lead to attackers attacking other people's devices through the network, and eventually lead to the or data breach

Images