Text matching optimization method and device
An optimization method and text technology, applied in neural learning methods, text database query, text database clustering/classification, etc., can solve the problems of slow express delivery, small intersection, why the express has not been sent out, etc., to improve the accuracy rate effect
Active Publication Date: 2021-06-29
北京沃丰时代数据科技有限公司
View PDF12 Cites 4 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
The text matching in the field of customer service is different from the general text matching. Some content that is considered unmatched in the general text expression can be matched in the field of customer service, such as: "Why hasn't the express delivery been delivered?" and "Express delivery The delivery is too slow", these two sentences both indicate that the customer is questioning the delivery speed of the express delivery, but from the words of the text, the intersection of these two sentences is very small, so the accuracy of text matching in the existing customer service field remains to be seen. further improvement
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0044] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention , but not all examples. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.
[0045] When matching text questions in the existing customer service field, it is mainly based on the pre-trained language model in natural language processing, and then tunes it according to specific tasks. Take the Bert model (Bidirectional Encoder Representations from Transformers) as an example, which is a Transformer bidirectional encoder designed to pre-train deep bidirectional represen...
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 provides a text matching optimization method and device. The method comprises the following steps: obtaining to-be-matched question text content; inputting the question text content into a trained text matching model, obtaining similar question texts of the question text content, and obtaining the trained text matching model by training a pre-trained language model through sample corpus data in the customer service field and a sample knowledge graph corresponding to the sample corpus data. According to the method, the training corpus and the knowledge graph are generated aiming at a specific customer service field environment, the text matching model is obtained through training by applying the deep learning model based on the training corpus and the knowledge graph, and the problem texts in the customer service field are matched, so that the text matching accuracy is improved.
Description
technical field [0001] The invention relates to the technical field of natural language processing, in particular to a text matching optimization method and device. Background technique [0002] In the intelligent question answering system in the field of customer service, it is necessary to match the text content of customer complaints or customer consultation with the standard question content stored in the existing text corpus knowledge base, so that according to the matching results, from these standard question content The answer scheme corresponding to the text content is obtained through analysis. [0003] When matching text questions in the existing customer service field, it is mainly based on the fine-tuning of the pre-trained language model in natural language processing to achieve text matching. This model is to learn the distributed representation of words, that is, the embedding process of words . This distributed representation will be used as an input and p...
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
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/33G06F16/332G06F16/35G06F16/36G06N3/04G06N3/08
CPCG06F16/3329G06F16/3346G06F16/367G06F16/35G06N3/08G06N3/045
Inventor 李伟
Owner 北京沃丰时代数据科技有限公司