Answer selection-oriented multi-angle attention feature matching method and system

A technology of feature matching and attention, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as complex feature engineering, lack of question information, lack of universality, etc., and achieve the effect of accurate and automatic acquisition

Active Publication Date: 2021-08-10
NANJING UNIV OF POSTS & TELECOMM
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the retrieval question answering system also faces the following problems: (1) The question information is scarce
[0005] (1) The method based on feature engineering requires artificial feature engineering. Such methods often have complex feature engineering, and the model can only adapt to specific data sets and is not universal.
Moreover, the effect of the model depends on the design of feature engineering, and the role of people is too important
[0006] (2) Among deep learning-based methods, representation-based methods usually perform relatively independent feature extraction on two sentences. Even if interaction information is added, it is a single-level interaction, and the interaction information is not fully utilized; interaction-based methods are usually difficult to consider Comprehensive interactive information, unable to consider local and global similarities and combine them effectively

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
  • Answer selection-oriented multi-angle attention feature matching method and system
  • Answer selection-oriented multi-angle attention feature matching method and system
  • Answer selection-oriented multi-angle attention feature matching method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0037] The specific implementation manners of the present invention will be further described in detail below in conjunction with the accompanying drawings.

[0038] The present invention relates to an answer selection-oriented multi-angle attention feature matching method. The best answer selection problem in the question answering system will inevitably face the problem of feature sparsity brought about by text characteristics. Considering that the traditional technology based on feature engineering method often has complex feature engineering, and the model can only adapt to a specific data set, it is not universal. The method proposed by the present invention has no such limitation, and through the use of the mixed coding layer, the coding information of the text is extracted more comprehensively and richly. Considering that when modeling the similarity between questions and answers, we often only pay attention to the similarity of a certain angle and cannot comprehensivel...

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 an answer selection-oriented multi-angle attention feature matching method and system, and the method comprises the steps: using a hybrid coding layer, enriching text features through employing the hybrid coding layer, extracting serialized information in combination with bidirectional LSTM, and obtaining wider short text information through employing convolution kernels of different heights of TextCNN, thereby achieving a complementary effect; using multi-level similarity calculation, extracting local interaction information and overall similarity features between two sentences by using multi-level similarity calculation and through an attention mechanism, then combining with the extracted features, and carrying out similarity scoring by considering local and overall sentences. In the technical scheme, from candidate answers, target answers can be automatically obtained efficiently and accurately.

Description

technical field [0001] The invention relates to an answer selection-oriented multi-angle attention feature matching method and system, belonging to the technical field of natural language processing. Background technique [0002] In the question answering system, after information retrieval, the content of the question answering system needs to sort the candidate answers so as to select the best answer. In scenarios such as community Q&A and intelligent customer service, it is more convenient and efficient to select the appropriate answer from the existing candidate answers than to generate the required answer. Answer selection is a very important task in natural language processing, and it is also an indispensable and important component of question answering systems. The answer selection task can be regarded as an answer matching task. Answer selection tasks are also considered as an application of text matching. However, the retrieval question answering system also fac...

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): G06F40/194G06F40/284G06K9/62G06N3/04G06N3/08
CPCG06F40/194G06F40/284G06N3/08G06N3/047G06N3/044G06N3/045G06F18/241
Inventor 徐小龙刘聪肖甫
Owner NANJING UNIV OF POSTS & TELECOMM
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