Unlock instant, AI-driven research and patent intelligence for your innovation.

Semantic parsing method and device based on rules and learning and electronic equipment

A semantic analysis and rule technology, applied in semantic analysis, machine learning, natural language translation, etc., can solve problems such as weak generalization ability, change, and difficulty in post-maintenance, to expand the scope of analysis, strong scalability, and solve limitations. sexual effect

Active Publication Date: 2021-02-09
北京智源人工智能研究院
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, these NLU tools have certain limitations. The analysis effect of NLU tools that fully use deep learning (such as DDParser and TexSmart) is affected by model training, and the model effect is unstable. For example, Textsmart cannot parse time periods into structured data. Unable to identify specific times represented by holidays, etc.
However, rule-based NLU tools (such as Duckling) can only recognize text within the scope of the rules, and cannot flexibly change with the development of the language. It is difficult to maintain later, and the generalization ability is weak.

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
  • Semantic parsing method and device based on rules and learning and electronic equipment
  • Semantic parsing method and device based on rules and learning and electronic equipment
  • Semantic parsing method and device based on rules and learning and electronic equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0042] Such as figure 1 As shown, the embodiment of the present invention provides a semantic analysis method based on rules and learning, including:

[0043] S101, identifying natural language sentences based on the original rules in the rule set, and generating structured sentences;

[0044] S102, judging whether the structured sentence can completely express the semantics of the natural language sentence, if not, obtaining the difference text between the natural language sentence and the structured sentence;

[0045] S103, inputting the difference text into a pre-trained learning model to generate new rules;

[0046] S104. Update the rule set by using the new rule.

[0047] In step S101, a large number of pre-written original rules are stored in the rule set, and the expression form of the original rules can be a self-defined context-free grammar rule. A syntax example might look like this:

[0048] =

[0049] |

[0050] |

[0051] |

[0052] |

[0053] |

...

Embodiment 2

[0104] Such as figure 2 As shown, another aspect of the present invention also includes a functional module architecture completely corresponding to the aforementioned method flow, that is, an embodiment of the present invention also provides a semantic analysis device based on rules and learning, including:

[0105] The rule parsing module 201 is used to identify natural language sentences based on the original rules in the rule set and generate structured sentences;

[0106] A judging module 202, configured to judge whether the structured sentence can completely express the semantics of the natural language sentence; if not, obtain the difference text between the natural language sentence and the structured sentence;

[0107] The new rule generating module 203 is used to input the difference text into the pre-trained learning model to generate new rules;

[0108] An update module 204, configured to update the rule set with the new rule.

[0109] The judgment ...

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 semantic parsing method and device based on rules and learning and electronic equipment. The method comprises the steps of recognizing natural language statements based on original rules in a rule set, and generating structured statements; judging whether the structured statements can completely express semantics of the natural language statement or not, if not, obtaininga difference text of the natural language statement and the structured statement, and inputting the difference text into a pre-trained learning model to generate a new rule; and updating the rule setby utilizing the new rule. By adopting the method provided by the invention, a new rule is obtained by utilizing machine learning, and the analysis range of a natural language is expanded, so that the continuously updated rule set not only can identify texts in the rule range, but also can be flexibly changed along with language development; and moreover, the method is high in expandability and generalization ability, and the problem of rule-based analysis limitation is solved.

Description

technical field [0001] The invention relates to the technical field of natural language understanding, in particular to a semantic analysis method, device and electronic equipment based on rules and learning. Background technique [0002] Natural Language Understanding (NLU, Natural Language Understanding) is a subfield of Natural Language Processing (NLP, Natural Language Processing), whose goal is to parse human language into machine-understandable and structured complete semantics. With the development of artificial intelligence technology, the enrichment of algorithms and the improvement of computing power, natural language understanding has also ushered in new developments in a new era. [0003] At present, NLU tools mainly include Baidu's DDParser, Tencent's TexSmart, Rasa NLU, Facebook's Duckling, and Google Syntaxnet. Among them, the first two use machine learning methods, and the latter three use rule-based recognition methods. However, these NLU tools have certai...

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): G06F40/30G06F40/194G06F40/56G06N20/00
CPCG06N20/00G06F40/194G06F40/30G06F40/56
Inventor 钱泓锦李晓桐刘占亮杨玉树窦志成曹岗文继荣
Owner 北京智源人工智能研究院