Asynchronous federal learning method and system based on block chain

Abstract

The invention provides an asynchronous federated learning method and system based on a block chain, and the method comprises the steps: introducing a block chain network on the basis of federated learning, and selecting an edge device node with a higher reputation as a conference node for the aggregation of a model and the reputation updating of all nodes, so as to achieve the decentralization of a federated learning process; and sending the global model to an edge device, training by using rich data generated by the edge device and returning updating parameters of the global model, respectively performing model aggregation, reputation updating and consensus authentication based on the negotiation node, and finally realizing complete federated learning under the condition of successful consensus. According to the method, in the model training process, data of edge equipment does not leave the local, the data leakage risk is reduced, and user privacy is protected; and based on decentration of a block chain network and a consensus algorithm, a single-point fault is avoided, and the robustness and expansibility of the system are improved.

Description

technical field [0001] The present invention relates to the technical field of electronic data processing, in particular to a blockchain-based asynchronous federated learning method and system. Background technique [0002] With the popularization and development of the Internet of Things and 4G / 5G wireless cellular network technology in the past ten years, the number of intelligent terminals at the edge of the network has increased significantly, and large-scale data collection and data interaction have led to data enrichment at the edge network. These massive terminal data can serve a wide range of artificial intelligence applications, enrich people's lives, and improve social productivity and work efficiency. By combining edge computing and AI technology, the local data sets of IoT devices can be directly utilized, and real-time decision-making and state awareness capabilities can be provided for IoT devices, which can better cope with complex real-world environments. Ho...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the embodiment of the present invention provides a block chain-based asynchronous federated learning method and system to eliminate or improve one or more defects existing in the prior art, and to solve the problems of single-point interaction in the federated learning process. Mistrust, data inconsistency, and problems with malicious nodes disrupting aggregates

Images