Deep learning-based question and answer matching method

A matching method and deep learning technology, applied in the field of question-answer matching based on deep learning, can solve problems such as heavy workload, low accuracy, and weak cross-domain, and achieve improved accuracy, good flexibility, and robustness , high efficiency effect

Active Publication Date: 2018-01-09
TONGJI UNIV
View PDF4 Cites 110 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] However, we found that the above existing methods mainly have obvious defects such as huge workload of feature engineering, weak cross-domain and low accuracy.

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
  • Deep learning-based question and answer matching method
  • Deep learning-based question and answer matching method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0015] Based on the technical solution summarized in the present invention, the technical solutions of the embodiments are further provided below, and the detailed process and details are as follows:

[0016] In step 1, for each pair of question and answer texts in the question-and-answer text set, the present invention first converts them into question word vectors and answer word vectors based on the Word2vec tool, thereby obtaining the question word vector matrix Q corresponding to the question-answer text set =(q 1 ,q 2 ,...,q l ) and answer word vector A=(a 1 ,a 2 ,...,a m ), where l and m are the number of questions and answers in the question-and-answer text set respectively, and q i (1≤i≤l) is the column vector corresponding to the i-th question, a j (1≤j≤m) is the column vector corresponding to the jth answer.

[0017] Then, the present invention inputs the question word vector matrix Q and the answer word vector A into the LSTM network to train and learn the s...

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 deep learning-based question and answer matching method. The method comprises the following steps of: 1) sufficiently learning word orders and sentence local features of a question text and an answer text by utilizing two underlying deep neural networks: a long short-term memory network LSTM and a convolutional neural network CNN; and 2) selecting a keyword with best semantic matching on the basis of a pooling manner of an attention mechanism AM. Compared with existing methods, the method has the advantages of being in low in feature engineering workload, strong in cross-field performance and relatively high in correctness, and can be effectively applied to the fields of commercial intelligent customer service robots, automatic driving, internet medical treatment, online forum and community question answering.

Description

technical field [0001] The present invention relates to the field of computer application technology, in particular to a question-answer matching technology based on deep learning. Background technique [0002] The intelligent question answering system mainly solves the real intention analysis of questions, the matching relationship between questions and answers, understands user questions described in natural language, and returns concise and accurate matching correct answers by searching heterogeneous corpora or question and answer knowledge bases . The processing framework of the question answering system includes three components: question understanding, information retrieval, and answer generation. According to the data domain of user questions, question answering systems can be divided into question answering systems for limited domains, question answering systems for open domains, and question answering systems for frequently asked questions (FAQ). The present inven...

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): G06F17/30G06N3/04
Inventor 黄震华
Owner TONGJI 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