Combinable weak authenticator-based named entity identification algorithm architecture

A technology of named entity recognition and entity recognition, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as large differences in entity boundaries, errors and failures in named entity recognition, and achieve increased robustness and enriched training samples Effect

Active Publication Date: 2021-04-23
SHANDONG UNIV
View PDF7 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In the prior art, when any one of the above two links makes a mistake, the final named entity recognition error will be caused: the current named entity recognition method usually regards the above two processes as one (traditional named entity recognition methods usually use the named entity Recognition is regarded as a sequence labeling task, using a deep network combined with a conditional random field for entity labeling, that is, combining entity boundary recognition and entity type recognition into one task), or processing one of the processes separately, resulting in the learning process, I do not know which one Links lead to t

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
  • Combinable weak authenticator-based named entity identification algorithm architecture
  • Combinable weak authenticator-based named entity identification algorithm architecture
  • Combinable weak authenticator-based named entity identification algorithm architecture

Examples

Experimental program
Comparison scheme
Effect test

Example Embodiment

[0127]Example 1

[0128]A algorithm architecture based on named entity identification based on a combined weak authenticator, including: entity identification part and result authentication section;

[0129]The entity identification portion is used to complete the identification task to obtain an identification result;

[0130]The result authentication portion includes two and more weak certificates, respectively, for verification of the identification result in each weak authenticator corresponding to the segmentation target.

[0131]The algorithm architecture also includes an information input layer, an entity identification layer, a data conversion layer, and a weak authenticator output layer;

[0132]The token information input layer is characterized by feature extraction: The layer includes a feature extraction module; after processing the text input and entity description input through the feature extraction module, as the first input information of the entity identification layer;

[0133]The ...

Example Embodiment

[0144]Example 2,

[0145]A algorithm architecture based on a nomenched entity identified as described in Example 1, the architecture further includes a training information flow, the training information flow specifically comprising: pre-training information flow and joint training information flow;

[0146]Among them, the pre-training information stream includes, entity identification module pre-training information flow, boundary weak certifier pre-training information flow and type weak certifier pre-training information;

[0147]Entity identification module pre-training information flow: The feature extraction module in the information input layer and the entity identification module in the entity identification layer participate in the training, the weak certification layer closes the input and output interface; the training data is the original training mark.

[0148]Boundary Weak Cautioner Module Predictive Information Flow: The Feature Extract Module in the Information Input layer and t...

Example Embodiment

[0151]Example 3,

[0152]As described in Example 1, an algorithm architecture based on a nomenched entity identified, the entity identification module includes a plurality of groups of neural networks, and the activation function of the neural network, preferably, each group of neural networks can be extracted sequences. The network structure of the feature, such as the Bi-LSTM neural network, Bi-GRU neural network, Bi-GRU neural network, or deep convolutional neural network.

[0153]Preferably, the feature extraction module is loaded with a pre-training language model based on self-focus mechanism; preferably, load the Bert algorithm.

[0154]The boundary weak authenticator module is loaded with neural networks (such as Bi-LSTM neural network, Bi-GRU neural network or deep convolutional neural network, etc.) and neural networks (such as SIGMOID or Softmax). ), Used to perform border legitimacy determination;

[0155]The type weak authenticator module is loaded with neural networks with extract...

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 combinable weak authenticator-based named entity identification algorithm architecture. The architecture comprises an entity identification part and a result authentication part, the entity identification part is used for completing an identification task to obtain an identification result; and the result authentication part comprises two or more weak authenticators which are respectively used for checking and authenticating the identification result on the subdivision target corresponding to each weak authenticator. The weak authenticator is a module capable of independently completing a subdivision target, and required training data can be automatically generated on an existing task corpus. The weak authenticator and the entity identification part form an end-to-end network and are used for carrying out optimization learning by using a supervision method. According to the invention, the combinable weak authenticator is used for assisting the named entity identification process, so that the entity identification precision is effectively improved, and the method can be simply and quickly adapted and expanded in an entity identification scene in a specific field.

Description

technical field [0001] The invention relates to a named entity recognition method and device based on a combinable weak authenticator, and belongs to the technical field of named entity recognition. Background technique [0002] Named entity recognition refers to the process of locating entity boundaries in text and classifying entities based on a predefined set of entity types. Named entity recognition results provide support for many downstream tasks such as knowledge graph construction, relation extraction, and information retrieval. Early named entity recognition mainly recognized simple entities such as names of people, places, and organizations. With the continuous expansion of the application field of named entity recognition, the types of entities have gradually increased, and there are some domain-specific entities in special fields. Type, such as the drug name in the biomedical field, etc. [0003] Named entity recognition can be subdivided into at least two proc...

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
IPC IPC(8): G06F40/295G06N3/04G06N3/08G06F40/30
CPCG06F40/295G06N3/049G06N3/08G06F40/30G06N3/044G06N3/045
Inventor 孙宇清吴佳琪刘天元
Owner SHANDONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap