Unlock instant, AI-driven research and patent intelligence for your innovation.
Short message template generation method and system based on deep learning, and electronic device
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
A SMS template, deep learning technology, applied in digital data processing, character and pattern recognition, instruments, etc., can solve problems such as adverse effects, users cannot accurately understand the meaning of SMS messages, users overdue loan repayments, etc., and achieve good results. Effect
Pending Publication Date: 2022-04-12
SICHUAN XW BANK CO LTD
View PDF0 Cites 0 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
The disadvantage of this scheme is that if it is only mechanized to delete words or sentences from the original text, key information in the text may be missed. When it is applied to the generation of SMS templates, it cannot accurately represent the meaning of SMS templates. In turn, the user cannot accurately understand the actual meaning of the text message, causing the user to repay the loan overdue or other adverse effects
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
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0060] This embodiment provides a method for generating short message templates based on deep learning, such as figure 1 shown, including the following steps:
[0061] Obtain the short message text to be compressed of the short message text, and preprocess the short message text to be compressed;
[0062] Wherein, the short message text to be compressed is the short message text after deleting the marked content, and the marked content is the content pre-marked in the short message text, and the marked content may be an important sentence marked manually.
[0063] Preprocessing is used to process the SMS text to be compressed into a form that can be recognized by the Transformer model. The traditional preprocessing method is word segmentation. The disadvantage is that in actual operation, if there are new words that have not appeared in the training data, then Unable to process. Therefore, in order to solve the above-mentioned problems, the present embodiment uses words as t...
Embodiment 2
[0085] This embodiment provides a short message template generation system based on deep learning, including:
[0086] The first obtaining module obtains the short message text to be compressed, and preprocesses the short message text to be compressed;
[0087] Wherein, the short message text to be compressed is the short message text after deleting the marked content, and the marked content is the content pre-marked in the short message text, and the marked content may be an important sentence marked manually.
[0088] Preprocessing is used to process the SMS text to be compressed into a form that can be recognized by the Transformer model. The traditional preprocessing method is word segmentation. The disadvantage is that in actual operation, if there are new words that have not appeared in the training data, then Unable to process. Therefore, in order to solve the above-mentioned problems, the present embodiment uses words as tokens for direct learning. In terms of specif...
Embodiment 3
[0107] This embodiment provides an electronic device, including a processor and a memory;
[0108] memory for storing processor-executable instructions;
[0109] The processor is configured to execute a method for generating short message templates based on deep learning as provided in Embodiment 1.
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
Login to View More
Abstract
The invention discloses a short message template generation method and system based on deep learning and an electronic device, and the method comprises the following steps: obtaining a to-be-compressed shortened message text of a short message text, and carrying out the preprocessing of the to-be-compressed shortened message text; inputting the preprocessed text to be compressed and shortened into a trained Transform model to obtain a compressed and shortened text; obtaining a statement smoothness probability matrix and a word frequency probability matrix of the compressed short letter text; and selecting an optimal compressed short letter text from the compressed short letter texts according to the statement smoothness probability matrix and the word frequency probability matrix. According to the short message template generation method and system based on deep learning and the electronic device, a large number of sentences similar to an original text are generated by building the abbreviation model, and then the sentences which conform to statement smoothness and have the number of words as small as possible are selected as output through the classification model, so that the purpose of generating the short message template is achieved.
Description
technical field [0001] The present invention relates to the technical field of natural languageprocessing in the computer field, in particular to a method, system and electronic device for generating short message templates based on deep learning. Background technique [0002] The current technology for text summarization can be divided into extractive summarization and generative summarization from the text source. The extractive summarization is to mark all the sentences in the article with weights, and extract some sentences as text summaries according to the weight. The main algorithms are as follows: : [0003] (1) TextRank algorithm, TextRank algorithm imitates the PageRankalgorithm and treats the sentences of text segmentation as nodes to form a hidden Markov chain, calculates the similarity between sentences to obtain the TextRank value of the text, and finally selects the sentences with the top TextRank scores as Thesis. [0004] (2) The text classification meth...
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
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to View More 
Login to View More