Event argument role extraction method based on multi-head attention mechanism

An attention and argument technology, applied in special data processing applications, instruments, unstructured text data retrieval, etc., can solve problems such as argument role extraction errors

Active Publication Date: 2019-08-16
温州开晨科技有限公司
View PDF8 Cites 63 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Aiming at the problem that there are errors in the extraction of argument roles when there are multiple events in a sentence in event

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
  • Event argument role extraction method based on multi-head attention mechanism
  • Event argument role extraction method based on multi-head attention mechanism
  • Event argument role extraction method based on multi-head attention mechanism

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0064] Obtain the event text to be extracted, and perform word segmentation processing on the text. After processing, the words are converted into word vectors through the trained word2vec model. Input the trained bidirectional GRU neural network to obtain the extracted event trigger words and corresponding event types.

[0065] Such as figure 1 , using the processed word vector and the corresponding label input to fuse the network of the multi-head supervisory attention mechanism, initialize the neural network parameters with the obtained weight, and adjust the neuron's weight through BP backpropagation according to the GRU part and the attention mechanism part. weights, to obtain a trained bidirectional GRU neural network with synergistic contextual attention. And use the trained two-way GRU model to process the event text to be extracted. Specific steps include:

[0066] Input the preprocessed training text (word vector and corresponding event annotation) into the bidir...

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 an event argument role extraction method based on a multi-head attention mechanism. The method comprises the following implementation steps: step (1), preprocessing a data settext, and outputting the preprocessed text and a corresponding tag; step (2), training a bidirectional GRU network integrated with a multi-head attention supervision mechanism; (3) performing semanticdependency analysis on the text, and outputting a semantic dependency path between the trigger word and the candidate argument; (4) inputting the preprocessed text into the network in the step (2) for training, outputting the code of each word, and fusing the semantic dependency path outputs (trigger word codes, candidate argument codes and semantic dependency paths) argument classification structures in the step (3); and (5) inputting the argument classification structure into a classification network for training and classifying. According to the method, the text is analyzed by using a neural network method fused with a multi-head attention supervision mechanism, and the method has a good argument role extraction capability for a plurality of event situations in event sentences.

Description

technical field [0001] The invention belongs to the technical field of natural language processing, relates to a method for classifying event argument roles, and provides an event argument role extraction method based on a multi-head attention mechanism. It is specifically used to extract event-related arguments from unstructured text and determine the role of arguments in the event. Background technique [0002] Event argument role extraction is to extract event-related information elements from unstructured information, and complete the event in a structured form. The current main research methods are pattern matching and machine learning. Pattern matching can achieve high performance in specific fields, but poor portability. Compared with pattern matching, machine learning has nothing to do with the domain, does not require too much guidance from domain experts, and has better system portability. With the construction of related corpora and the continuous enrichment of...

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): G06F16/31G06F17/27G06F16/35
CPCG06F16/313G06F16/353G06F40/30
Inventor 汤景凡戚铖杰张旻姜明闻涛
Owner 温州开晨科技有限公司
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