Federal learning system and federal learning training method for smart city Internet of Things belief fusion

Abstract

The invention discloses a federated learning system oriented to belief fusion of the Internet of Things of a smart city. The federated learning system comprises an equipment end and a server end, the equipment end comprises an MQTT, a training thread and a heartbeat sending thread; the server comprises an MQTT, a model aggregation thread, a task scheduling thread and a heartbeat monitoring thread. The invention also provides a federated learning training method based on the federated learning system, and the method comprises the steps: the server initializes a global model, selects a certain proportion of equipment in a communication round, and sends the global model to the selected equipment; after the equipment receives the message, the model is trained and updated; the updated model is uploaded to the server again and aggregated, a new global model is obtained, and the next communication round is entered; and after the preset number of communication rounds is completed, the device receives the latest global model issued by the server and stores the latest global model for subsequent use. The method has certain equipment expandability, robustness and high efficiency.

Description

technical field [0001] The invention belongs to the technical field of Internet of Things communication technology, deep learning algorithm, federated learning algorithm and information fusion technology in the Internet of Things scenario, and relates to the use of Internet of Things communication technology to realize the interconnection of multiple device clients and servers, and to build a communication interface for federated learning model transmission. , to carry out federated learning with the participation of multi-party devices in the smart city IoT scenario, and to realize the federated learning system and federated learning training method of the integration of trust and matter. Background technique [0002] With the advancement of mobile technology, edge devices such as the Internet of Things and smart mobile phones have achieved rapid development and become an indispensable part of modern life. These devices have certain computing and communication capabilities ...

AI Technical Summary

Problems solved by technology

In addition, in the actual scenario of the Internet of Things in a smart city, the local data of each participating device is often non-independent and identically distributed (Non-IID) and unbalanced, because the performance of different devices is different, and the scenarios they are in are also different. The collected data will have a large gap in quantity and distribution, and the heterogeneity of this data will cause the client to introduce offsets in the process of performing local updates, making the local model deviate from the global model, resulting in "client drift problem", which in turn leads to slow and unstable convergence process, impairing the performance of the model

Images