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

Question Entity Recognition and Linking Method, Device, Computer Equipment and Storage Medium

A technology of entity recognition and entity, which is applied in the field of knowledge graph, can solve the problems of limited generalization ability of neural network model, influence of entity link, poor use of entity hierarchical structure, etc., and achieve the effect of reducing error transfer effect and reasonable entity link

Active Publication Date: 2022-05-17
PING AN TECH (SHENZHEN) CO LTD
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the generalization ability of the neural network model is limited. In practical applications, the NER model is likely to mistake the start and end positions of an entity in the question, which has a great impact on subsequent entity links.
In addition, most of the current knowledge graph question answering systems are in the open field. Due to too many types of entities, the entity hierarchy structure cannot be well utilized.
However, in verticals, the hierarchy of entities is clearer

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
  • Question Entity Recognition and Linking Method, Device, Computer Equipment and Storage Medium
  • Question Entity Recognition and Linking Method, Device, Computer Equipment and Storage Medium
  • Question Entity Recognition and Linking Method, Device, Computer Equipment and Storage Medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0038] see figure 1 , a method for entity recognition and linking of questions in the present embodiment, comprising the following steps:

[0039] S1 receives the user question, identifies the entity to be confirmed in the user question, and the preliminary entity boundary and entity category of the entity to be confirmed, and the preliminary entity boundary is the preliminary entity boundary of the entity to be confirmed in the user question start position and preliminary end position;

[0040]In this step, the accuracy of entity boundary identification is not high, and only the preliminary position of the entity boundary in the user sentence needs to be identified for subsequent entity matching confirmation. In a specific implementation manner of this embodiment, entities, entity boundaries, and entity categories in user sentences are identified based on a named entity recognition model (Named Entity Recognition, NER). NER (also known as entity recognition, entity block an...

Embodiment 2

[0058] read on figure 2 , the present invention shows a question entity recognition and linking device. In this embodiment, the question entity recognition and linking device 10 may include or be divided into one or more program modules, and one or more program modules are divided into It is stored in a storage medium and executed by one or more processors to complete the present invention and realize the above-mentioned question entity recognition and linking method. The program module referred to in the present invention refers to a series of computer program instruction segments capable of completing specific functions, which is more suitable than the program itself to describe the execution process of the question entity recognition and linking device 10 in the storage medium. The following description will specifically introduce the functions of each program module of the present embodiment:

[0059] The confirmation entity identification module 11 is used to receive us...

Embodiment 3

[0073]The present invention also provides a computer device, such as a smart phone, a tablet computer, a notebook computer, a desktop computer, a rack server, a blade server, a tower server or a cabinet server (including an independent server, or a multi- A server cluster composed of servers), etc. The computer device 20 of this embodiment at least includes but is not limited to: a memory 21 and a processor 22 that can be connected and linked through a system bus, such as image 3 shown. It should be pointed out that, image 3 Only computer device 20 is shown having components 21-22, but it should be understood that implementing all of the illustrated components is not a requirement and that more or fewer components may instead be implemented.

[0074] In this embodiment, the memory 21 (that is, a readable storage medium) includes a flash memory, a hard disk, a multimedia card, a card-type memory (for example, SD or DX memory, etc.), a random access memory (RAM), a static ra...

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 question entity identification and linking method, which includes receiving user questions, identifying entities to be confirmed, preliminary entity boundaries and entity categories in the user questions, adjusting the position numbers of the preliminary start position and the preliminary end position, Obtain multiple adjusted entity boundaries as the candidate entity boundaries of the entity to be determined; determine the candidate entities in the user question according to the candidate entity boundaries; search each candidate entity in the knowledge map according to the determined entity category, and determine each The candidate entity name corresponding to the candidate entity, and determine the similarity between the candidate entity and the candidate entity name according to the similarity algorithm; select the candidate entity name with the highest similarity and the corresponding candidate entity boundary as the result of entity linking and output. It can effectively reduce the error transfer effect of the NER model on the subsequent steps of the question answering system, and at the same time make reasonable use of the map hierarchy to perform more reasonable entity links.

Description

technical field [0001] The present invention relates to the field of knowledge graphs, in particular to a question entity recognition and linking method, device, computer equipment and storage medium. Background technique [0002] At present, the bi-LSTM-CRF network is widely used in the industry and academia to perform various entity recognition (NER) tasks, including entity recognition in knowledge base question answering. However, the generalization ability of the neural network model is limited. In practical applications, the NER model is likely to mistake the start and end positions of an entity in the question, which has a great impact on subsequent entity links. In addition, most of the current knowledge graph question answering systems are in the open field. Because there are too many types of entities, they cannot make good use of the entity hierarchy. However, in the vertical domain, the hierarchy of entities is clearer. For example, in the knowledge map of the m...

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 Patents(China)
IPC IPC(8): G06F40/295
CPCG06F16/00
Inventor 朱威李恬静
Owner PING AN TECH (SHENZHEN) CO LTD