Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A method for missing semantic complementation in multi-turn question answering systems

A question-answering system and technology with missing semantics, applied in the field of information processing, can solve problems such as the inability to find the correct answer for sentences with missing semantics, and achieve the effect of improving fluency and accuracy and good user experience

Active Publication Date: 2017-08-08
北京中科汇联科技股份有限公司
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is how to overcome the defect that the correct answer cannot be found for sentences with missing semantics in the multi-round question answering system in the prior art

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
  • A method for missing semantic complementation in multi-turn question answering systems
  • A method for missing semantic complementation in multi-turn question answering systems
  • A method for missing semantic complementation in multi-turn question answering systems

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0078] Example 1: "What business does Haidian District Customs have?", "What is its phone number?"

[0079] Construction of expression pairs: According to the recognition results of referring pronouns and candidate antecedents—the referring pronoun set and the candidate antecedent set, Cartesian product operation is performed on the two sets to obtain a set of expression pairs, such as in Example 1, the set of candidate antecedents is {"Haidian District Customs", "business"}, the set of pronouns is {"it"}, and the calculated expression pair set is {"it"-"Haidian District Customs", "it"-"business"} .

[0080] Generation of expression pair features: The expression pair features are composed of three aspects, one is artificial features, the other is word vector features, and the third is interaction features. The artificial features are artificially proposed and have practical meanings, including the antecedent features "person", "Male", "Female", "Singular", "Plural", "Item", "...

example 2

[0090] Example 2: How to handle bank card transfer in different places and different banks?

[0091] c. Model omission recovery: The method of omission recovery is to use the nouns or verbs that appear above as candidate words, calculate the co-occurrence probability p of the candidate word and the word in the current sentence, and set the threshold d. For the word w, its co-occurrence probability p>threshold d, and the word w does not appear in the current sentence, the word w is used as a semantic supplementary word for ellipsis recovery, as in example 3, according to the model trained by b, assuming that the model only trained the sentence of example 2, above After Ansj processing, the candidate target words are "Bank of China", "provide", "bank card", "transfer" and "service", and the co-occurrence probability with the current sentence "handling" is p (handling, Bank of China)=0.0, p(handle, provide)=0.0, p(handle, bank card)=0.0, p(handle, transfer)=0.0, p(handle, service...

example 3

[0092] Example 3: "Does Bank of China provide bank card transfer service?", "How to handle it?"

[0093] Finally, because both the reference resolution module and the omission recovery module use the Ansj natural language processing tool, in order to improve the operating efficiency, the design is as follows: Figure 5 , as shown in the overall block diagram, put the common operation part before the module, set up an independent preprocessing module, and transmit the processing results to the reference resolution module and the omission recovery module respectively, and set up a post-processing module, which will refer to the resolution module and The processing results of the omission recovery module are merged together and output to the problem retrieval system.

[0094] The anaphora resolution module is based on the improvement of the model by introducing word vector features, which can effectively capture the contextual semantic features of sentences. At the same time, the...

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 missing semantic supplementing method for a multi-round question-answering system. The missing semantic supplementing method is characterized by comprising the following steps of S1, obtaining questions input in the question-answering system by a user; S2, performing missing semantic supplementation for the current questions according to previous questions input by the user, wherein the missing semantic supplementation refers to semantic supplementation from the point of anaphora resolution and / or omission and recovery; and S3, performing searching on the current questions after the missing semantic supplementation is completed. According to the missing semantic supplementing method, the missing semantic supplementation for the current questions is carried out through the anaphora resolution and / or omission and recovery; accurate answers for the questions input by the user can be obtained; and the method allows the user to ask questions by using elliptical sentences, so that the fluency and accuracy of the man-machine interaction are improved, and a better user experience is achieved.

Description

technical field [0001] The invention relates to the technical field of information processing, in particular to a method for supplementing missing semantics in a multi-round question answering system. Background technique [0002] Automatic question answering system, also known as QA (Question Answering) system, can be called a new generation of search engine. Users do not need to decompose their questions into keywords, but can directly submit the whole question to the system, and can use natural language sentences Asking questions can also directly return answers for users, which can better meet users' retrieval needs. Therefore, automatic question answering systems have become the first choice for enterprise intelligent customer service, but there are still some shortcomings in this system. It is independent and cannot establish an accurate context. Therefore, for users, each search must enter a question with complete semantics, which is contrary to human language behavio...

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 Patents(China)
IPC IPC(8): G06F17/27G06F17/30
CPCG06F16/3329G06F40/30
Inventor 游世学杜新凯
Owner 北京中科汇联科技股份有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com