Recommendation method, device and equipment for personalized content

Abstract

The present application relates to a method, device and equipment for recommending personalized content. Among them, a personalized content recommendation method is applied to the server, and after receiving the content request from the client for the target content, multiple candidate contents corresponding to the target content are obtained; and the current reinforcement learning model to determine the target model; among them, the standby model is an algorithm used to determine personalized content based on multiple candidate contents, and is in the online use state when the current reinforcement learning model is online; the preset calling rules are used to ensure The first probability of selecting the current reinforcement learning model as the target model within the first preset time period is less than the second probability of selecting the backup model; using the target model to determine the target content from a plurality of candidate content, recommending the target to the client content. This solution can improve the accuracy of the target content recommended to users within a certain period of time when the reinforcement learning model is online.

Description

technical field [0001] The present application relates to the technical field of machine learning, and in particular to a method, device and equipment for recommending personalized content. Background technique [0002] In order to improve user experience, various applications, such as music applications, shopping applications, news applications, and video applications, etc., will recommend personalized content according to the user's preferences for different users. In related technologies, after receiving the content request sent by the client for obtaining the target content, the servers of various applications obtain multiple candidate contents corresponding to the target content, and then use a reinforcement learning model to determine from the multiple candidate contents The target content is the content to be recommended to the user, so as to recommend the determined target content to the client. [0003] The above-mentioned reinforcement learning model can be obtain...

AI Technical Summary

Problems solved by technology

[0004] However, during online training, it still takes a certain amount of time for the reinforcement learning model to be trained to accurately determine the target content. In this way, the accuracy of the target content recommended to the user is low within a certain period of time after starting the training.

Images