Information recommendation method and device and storage medium

Abstract

The invention provides an information recommendation method and device, and a storage medium, relates to the cloud technology, relates to the technical field of computers, and is used for improving the scoring precision. In the present application, the method comprises: after receiving an information scoring request, sending a request message carrying a scoring service model identifier to an online parameter server so as to extract model parameters and a model network structure of the scoring service model from the online parameter server, the model parameters and the model network structure of the trained scoring service model being distributed and stored in the online parameter server; determining the trained scoring service model based on the model parameters and the model network structure; and scoring the to-be-recommended information in the information scoring request based on the trained scoring service model, and returning an obtained scoring result to the upstream server, so that the upstream server recommends the to-be-recommended information to the target user according to the scoring result.

Description

technical field [0001] The present application relates to the field of computer technology, and provides an information recommendation method, device and storage medium. Background technique [0002] For information recommendation, the machine learning model is used to score information, sort the information according to the scoring results, and recommend information to users according to the ranking, so that users can obtain the information they are interested in. Therefore, in the process of information recommendation, it is necessary to train various machine learning models for the sorting and recalling links of the information recommendation system. Among them, the core of the two links of sorting and recall is to score information, which is the most important scoring service model for information recommendation. [0003] At present, all parameters of the scoring service model need to be loaded into the stand-alone service, that is, loaded into the content scoring serve...

AI Technical Summary

Problems solved by technology

[0003] At present, all parameters of the scoring service model need to be loaded into the stand-alone service, that is, loaded into the content scoring server, but when all the parameters of the model reach the upper memory limit of the content scoring server, the parameters of the scoring service model cannot be loaded to the content scoring server , causing the content scoring server to fail to score the information;

[0004] And the model parameters and model network structure of the new scoring service model generated by each training will change, and the new scoring service model generated by training will be uploaded to the online service in the form of a model file, and the online service will download the model file and obtain The model parameters and model network structure in the model file are merged to generate a model. It takes time to store and download the model file in the network, so there is a certain delay in the update of the model, resulting in a decrease in scoring accuracy

Images