Unlock instant, AI-driven research and patent intelligence for your innovation.

Recommendation model training method and device

A model training and model technology, applied in character and pattern recognition, prediction, instrumentation, etc., can solve problems such as high time complexity

Pending Publication Date: 2020-11-17
HUAWEI TECH CO LTD
View PDF0 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This process requires step-by-step training, and the time complexity is high

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Recommendation model training method and device
  • Recommendation model training method and device
  • Recommendation model training method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 2

[0138] The recommendation model provided in Embodiment 2 is obtained through comprehensive training based on the second training set of at least two users and the comprehensive loss function after being trained according to the first training set of each of the at least two users. Predicting the recommendation probability of the recommended object of the user to be predicted according to the comprehensively trained recommendation model can meet the individual needs of the user to be predicted and improve the prediction accuracy.

[0139] See Image 6 , Embodiment 3 of the present application also provides an apparatus 30 for recommending model training. The device 30 for training a recommended model is used to implement the method for training a recommended model provided in Embodiment 1 of the present application. The device 30 for recommending model training may include an acquisition module 301 , an inner training module 302 , an outer training module 303 and an online pre...

Embodiment 4

[0168] In the device provided in Embodiment 4, the recommendation model in the online prediction module is trained according to the first training set of each of at least two users, and then based on the second training set of at least two users and the comprehensive loss obtained through function synthesis training. Predicting the recommendation probability of the recommended object of the user to be predicted according to the comprehensively trained recommendation model can meet the individual needs of the user to be predicted and improve the prediction accuracy.

[0169] Figure 8 It is a schematic structural diagram of a recommended model training device provided in this application. like Figure 8 As shown, the device 50 includes a processor 501 , a memory 502 and a transceiver 503 , and the processor 501 , the memory 502 and the transceiver 503 may be connected through a bus 504 .

[0170] The device 50 is a device with a hardware structure, which can be used as Ima...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a recommendation model training method and device, and the method comprises the steps: obtaining user sample data of at least two users, wherein user sample data of each user in the at least two users comprises a first training set and a second training set; training a recommendation model according to the first training set of each user, with the recommendation model comprising K recommendation list models and a model selector, and K being an integer greater than 1; and obtaining model parameters of the trained recommendation model of each user; comprehensively training a recommendation model according to second training sets of the at least two users and a comprehensive loss function to acquire the recommendation model after comprehensive training, with the comprehensive loss function being obtained according to model parameters of the recommendation model of each user. The method does not need to manually define meta-features. Meanwhile, a recommendation model comprises K recommendation single models and a model selector, the K recommendation single models and the model selector can be jointly trained and subjected to parameter updating, step-by-step training is not needed, and time complexity can be reduced.

Description

technical field [0001] The present invention relates to the technical field of content recommendation, in particular to a method and device for training a recommendation model. Background technique [0002] With the development of Internet technology and the rapid growth of information, how to screen information quickly and effectively, so as to accurately recommend personalized content suitable for users (such as commodities, advertisements, news information, videos, music, reading, applications, etc.) To users is an important research topic at present. Recommending personalized content to users is an important means of monetizing artificial intelligence and a sharp tool to improve user experience and platform revenue. [0003] Since the data of different users is very different, it is difficult for a recommendation model to apply to all users. See figure 1 , figure 1 Recommend-to-monetization curve graph over the course of a day for four recommendation models. figure ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06F16/9535G06Q10/04
CPCG06F16/9535G06Q10/04G06F18/214
Inventor 程朋祥陈飞董振华李震国何秀强
Owner HUAWEI TECH CO LTD