Remote supervision relation classification method based on PCNN and multi-layer attention

A technology of relationship classification and remote supervision, applied in the information field, can solve problems such as inability to model training and small amount of data, and achieve the effects of reducing manpower, making full use of it, and reducing noise

Active Publication Date: 2019-12-10
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF9 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009]The data set is too closed, the available data is only manually labeled data, the amount of data is small, and the model cannot be trained well

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
  • Remote supervision relation classification method based on PCNN and multi-layer attention
  • Remote supervision relation classification method based on PCNN and multi-layer attention
  • Remote supervision relation classification method based on PCNN and multi-layer attention

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0077] The technical solutions of the present invention will be further described below in conjunction with the drawings and specific embodiments.

[0078] Such as figure 1 As shown, a multi-instance learning remote supervised relation classification method based on PCNN and Multi-level attention, including two stages, training relation classification model stage and prediction stage;

[0079] (1) Training relational classification model stage

[0080] Step 1: Preprocess the training corpus of the relational classification model

[0081] (1) Convert the original corpus data OrgData into character-level corpus data NewData;

[0082] (2) count the characters of NewData, obtain the character set CharSet, number each character, and obtain the corresponding character number set CharID of the character set;

[0083] (3) Convert the entities, relations, and sentences in each bag through CharID to obtain the bag expressed in the form of ID;

[0084] (4) Obtain the list representat...

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 relates to a remote supervision relation classification method based on PCNN and multi-layer attention, and belongs to the technical field of information. The method comprises two stages: a stage of training a relationship classification model and a prediction stage. The stage of training the relationship classification model further comprises the following steps: 1) preprocessing training corpora of the relationship classification model; 2) establishing a relationship classification model based on the PCNN and Multi-level action, and training model parameters; 3) training a model and optimizing model parameters; 4) storing the trained relationship classification model. In the prediction stage, the trained relation classification model is used for carrying out relation category prediction on to-be-predicted data. According to the algorithm provided by the invention, the relationship type between entities in the sentence can be determined more accurately, and a good foundation is provided for some downstream work; many manpower for data annotation is reduced, and high accuracy is achieved.

Description

technical field [0001] The invention relates to the field of information technology, in particular to a remote supervision relationship classification method based on PCNN and multi-layer attention. Background technique [0002] Traditional relation classification methods can be divided into rule-based relation classification methods, relation classification methods based on traditional machine learning, deep learning relation classification methods based on full supervision, and deep learning relation classification methods based on remote supervision. The method of rule-based relationship classification is based on the principle of using hand-written rules to match text and rules to classify relationships. For example, in the found(PERSON, ORGANIZATION) mode, if PERSON and ORGANIZATION are included in the text, the relationship between the two entities is considered to be a founded relationship. The method of relationship classification based on traditional machine learni...

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): G06F16/28G06F16/2455G06K9/62G06N3/04
CPCG06F16/288G06F16/24564G06F16/285G06N3/045G06F18/214
Inventor 廖伟智叶光磊马亚恒左东舟
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
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