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

Intention understanding method and device

An intent and generalization technology, applied in the field of information processing, can solve the problem of heavy manual labeling workload, and achieve the effect of reducing the workload and difficulty of labeling

Pending Publication Date: 2020-01-10
BEIJING BAIDU NETCOM SCI & TECH CO LTD
View PDF3 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] In related technologies, the oral dialogue comprehension task uses training data such as labeled samples or templates to realize the recognition of the intent and entity words of spoken sentences, matches based on manually labeled templates, and uses a certain entity recognition method to label word slot labels. The natural language that meets the template conditions returns the corresponding intent and entity word recognition results. However, this method of intent understanding relies on artificial abstraction to summarize the laws of natural language, so as to mark a large number of high-quality templates to achieve template coverage of the business scope. , leading to a large workload of manual labeling

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
  • Intention understanding method and device
  • Intention understanding method and device
  • Intention understanding method and device

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0051] In this example, a generalization tree structure is constructed in advance, and the target generalization text is matched with multiple generalization templates based on the generalization tree structure.

[0052] That is, multiple template texts are acquired, and then, template word slot tags corresponding to each template text constituent word in the multiple template texts are identified, and further, generalized templates of all template texts are generated according to the template word slot tags.

[0053] For example, if Figure 4 As shown, the template text is: Book a meeting room in Building No. 3 at 2:00 pm today, and the recognized results are: 2:00 pm today (user_time), Building 3 (user_location), and the query contains two word slots The words corresponding to the tags, the first one is "2 o'clock in the afternoon today", which is the time when the meeting room is scheduled, expressed as user_time, and the second is "Building No. 3", which is the location of...

example 2

[0058] In this example, in order to improve the matching efficiency, the number of target word slot labels contained in the target generalization text is obtained, and among the multiple pre-trained generalization templates, the generalization template containing the same number of word slot labels is obtained as a candidate Generalized templates.

example 3

[0060] In this example, the target word slot tags contained in the target generalized text are obtained, and the generalization templates containing the same word slot tags as the target word slot tags exceed a certain value are used as candidate generalization templates.

[0061] Step 103, calculating the semantic similarity between the target generalization text and the candidate generalization templates, and determining the target generalization template among the candidate generalization templates according to the semantic similarity.

[0062] It is easy to understand that the candidate generalization template is only a generalization template that is preliminarily screened out and similar to the target generalization text. Therefore, it is necessary to further calculate the semantic similarity between the target generalization text and the candidate generalization template. The target generalization template is determined in the generalization template.

[0063] It should...

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 provides an intention understanding method and device, and the method comprises the steps: recognizing a target word slot label corresponding to a composition word in a target text, andgenerating a target generalized text of the target text according to the target word slot label; matching the target generalization text with a plurality of preset generalization templates, and determining candidate generalization templates according to the matching degree; calculating the semantic similarity between the target generalization text and the candidate generalization template, and determining the target generalization template according to the semantic similarity; and obtaining a template intention of the target generalization template, taking the template intention of the targetgeneralization template, and generating an intention understanding result of the target text according to the template intention, the target word slot label and the composition words corresponding tothe target word slot label. Therefore, intention understanding of related information is realized based on a generalization processing mode, dependence on a large number of annotation training samplesis avoided, annotation workload and difficulty of a sample person are reduced, and intention understanding modes are enriched.

Description

technical field [0001] The present invention relates to the technical field of information processing, in particular to an intent understanding method and device. Background technique [0002] Usually, understanding the intention of natural language is an indispensable part of the current intelligent scene. The task of understanding spoken dialogue is one of the tasks of natural language processing. The usual processing method is: define an intention set and entity word set, Each intent can correspond to a subset of the entity word set, and each spoken sentence can correspond to one or more intents in the intent set, and the sentence can contain several word fragments, and each fragment corresponds to an entity word. [0003] In related technologies, the oral dialogue comprehension task uses training data such as labeled samples or templates to realize the recognition of the intent and entity words of spoken sentences, matches based on manually labeled templates, and uses a ...

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): G06F16/33
CPCG06F16/3344
Inventor 于博
Owner BEIJING BAIDU NETCOM SCI & TECH CO LTD
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