Decentralized federal recommendation method, intelligent terminal and storage medium

Abstract

The invention discloses a decentralized federated recommendation method and device, an intelligent terminal and a storage medium, and the decentralized federated recommendation method comprises the steps: obtaining actually scored articles to construct a first article set, and randomly sampling and obtaining unscored articles to construct a second article set; combining the first article set and the second article set into a third article set, and sending the third article set to a server; on the basis of the third article set, a communication matrix is constructed through a server, and the communication matrix is used for recording whether all the articles are scored by all the users or not; and based on the third article set, obtaining a user set corresponding to each article scored together in the communication matrix and using the user set for model training. According to the scheme, user privacy can be protected through decentralized federal recommendation, and a good recommendation effect, namely recommendation accuracy, is achieved.

Description

technical field [0001] The invention relates to the technical field of federated recommendation, in particular to a decentralized federated recommendation method, an intelligent terminal and a storage medium. Background technique [0002] With the improvement of people's awareness of privacy protection, traditional methods such as collaborative filtering algorithms that collect user data to the server for modeling and information recommendation are no longer feasible because they will leak user privacy. In order to achieve the purpose of protecting user privacy and completing information recommendation, people try to combine the recommendation system with federated learning, that is, federated recommendation. [0003] In recent years, many centralized federated recommendation algorithms have been proposed, such as federated collaborative filtering method FCF and explicit feedback federated recommendation method FedRec. These algorithms require the participation of servers to...

AI Technical Summary

Problems solved by technology

[0007] Aiming at the technical problems in the prior art, such as large model training deviation, high communication cost, possible disclosure of user scoring behavior and poor recommendation performance in the centralized federated algorithm and the decentralized federated algorithm, the present invention provides a de-centralized federated algorithm. Centralized federated recommendation method, intelligent terminal and storage medium, including a sharing strategy and two corresponding updating strategies

Images