Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Searching-ranking model training method and device and search processing method

A sorting model and training method technology, applied in the direction of electronic digital data processing, special data processing applications, network data retrieval, etc., can solve the problem of failing to examine the relationship between words and words, not investigating, and not considering polysemy , synonyms and other issues to achieve accurate search and sort results and improve accuracy

Active Publication Date: 2015-05-13
BAIDU ONLINE NETWORK TECH (BEIJIBG) CO LTD
View PDF6 Cites 56 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although this method solves the problem that the traditional method does not consider polysemy and synonyms to a certain extent, however, the vector representation of the sentence formed by simply adding word vectors fails to examine the relationship between words. Dependencies, for example, the simple addition of word vectors in the search term "Jackie Chan's Fame Works" fails to consider that both "Jackie Chan" and "Fame" are modifiers of the word "Works"
Therefore, the ranking scores obtained based on this method are not very accurate

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
  • Searching-ranking model training method and device and search processing method
  • Searching-ranking model training method and device and search processing method
  • Searching-ranking model training method and device and search processing method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0023] figure 2 It is a flow chart showing the training method of the Gated RNN-based search ranking model in Embodiment 1 of the present invention. refer to figure 2 , the training method of the search ranking model based on Gated RNN comprises the steps:

[0024] In step S110, multiple sets of labeled sample data are acquired, each set of sample data includes a search term and its corresponding search result items marked as positive or negative.

[0025] According to the concept of the present invention, the search result items marked as positive examples are search result items that have been clicked, and the search result items marked as negative examples are search result items that have not been clicked. Specifically, when the user inputs a search term, multiple search result items will be obtained, and the user selects one of the search result items for further browsing, and the selected search result item is the clicked search result item. Otherwise, it is a searc...

Embodiment 2

[0054] Image 6 It is a flow chart showing the search processing method in Embodiment 2 of the present invention. refer to Image 6 , the method can be executed, for example, on a search engine server. The search processing method includes the following steps:

[0055] In step S210, a user's search term is received.

[0056] The search term may be a search term sent from a client. For example, the user inputs "vehicle violation query" on the browser search engine interface to search, and the browser application sends the search term to the search engine server.

[0057] In step S220, a plurality of search result items are acquired according to the search term.

[0058] A search engine server may retrieve a plurality of search result items using existing search techniques (eg, from a pre-compiled index of web pages) using the search term.

[0059] In step S230, with the search word and the plurality of search result items as input, the ranking score of each ...

Embodiment 3

[0065] Figure 7 It is a logical block diagram showing the training device of the Gated RNN-based search ranking model according to Embodiment 3 of the present invention. refer to Figure 7 , the training device of the Gated RNN-based search ranking model includes a sample data acquisition module 310 , a search ranking model generation module 320 and a parameter learning module 330 .

[0066] The sample data acquisition module 310 is used to acquire multiple sets of labeled sample data, each set of sample data includes a search term and its corresponding search result items marked as positive or negative.

[0067] Preferably, the search result items marked as positive examples are search result items that have been clicked, and the search result items marked as negative examples are search result items that have not been clicked.

[0068] The search ranking model generation module 320 is used to generate the input layer, word vector layer, hidden layer and output l...

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

An embodiment of the invention provides a searching-ranking model training method and device and a search processing method. The searching-ranking model training method includes: acquiring multiple groups of labeled sample data, wherein each group of the sample data includes search terms and multiple search result entries correspondingly labeled as positive examples or negative examples; generating Gated RNN (recurrent neural network) based input layer, word vector layer, hide layer and output layer of a searching-ranking model, and training the searching-ranking model to learn parameters of the same. By the searching-ranking model training method and device and the search processing method, accuracy of ranking scores of the search result entries acquired by calculating can be improved, and more accurate searching-ranking results are provided for users.

Description

technical field [0001] The invention relates to the technical field of natural language processing, in particular to a training method for a search ranking model, a search processing method and a device. Background technique [0002] With the development of Internet applications, search processing technology is becoming more and more mature. It is a search engine system that expresses the search term entered by the user in a specific form to calculate the ranking score with the search result items to be sorted (take the title of the webpage as an example), so as to return more accurate search ranking results according to the ranking score core issue. [0003] The traditional method of calculating the ranking score is to calculate the degree of similarity between the two text strings as a sorting score by calculating the exact match between the search term and the words contained in the search result item (taking the title of the webpage as an example). Fraction. For examp...

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): G06F17/30
CPCG06F16/90348G06F16/9038G06F16/951
Inventor 张军
Owner BAIDU ONLINE NETWORK TECH (BEIJIBG) CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com