An end-to-end intent recognition system and method based on memory forgetting device

A technology for identifying systems and intentions, applied in the computer field, can solve problems such as errors in recognition results and inability to cover sentences well, and achieve the effect of ensuring lossless transmission

Active Publication Date: 2020-06-16
CHENGDU XIAODUO TECH CO LTD
View PDF18 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In this scheme, LSTM is based on vocabulary-level memory and forgetting, which cannot cover sentences very well, and only pays attention to the relationship between the current question and the above sentence. When all the above questions have nothing to do with the current question, it will Cause recognition error

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
  • An end-to-end intent recognition system and method based on memory forgetting device
  • An end-to-end intent recognition system and method based on memory forgetting device
  • An end-to-end intent recognition system and method based on memory forgetting device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0034] The present invention will be further described below in conjunction with accompanying drawing:

[0035] Such as figure 1 As shown, an end-to-end intent recognition system based on a memory forgetting device includes: a data preprocessing unit, a contextual intent recognition unit, a single sentence recognition unit, and a contextual intent determination unit.

[0036] The data preprocessing unit performs sentence-level vector representation on the input sentence to obtain the sentence vector of each sentence. The input sentence includes several preceding and current question sentences. The data preprocessing unit includes a sentence numerical conversion unit and a sentence vectorization unit, and the sentence numerical conversion unit performs numerical conversion on the sentence to obtain a vector of a specified length, and forms a first matrix of multiple vectors in the same data set ; The sentence vectorization unit uses the pre-training model to calculate the mat...

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 an end-to-end intent recognition system and method based on a memory forgetting device, which includes a data preprocessing unit, which performs sentence-level vector representation on input sentences to obtain sentence vectors for each sentence; the input sentences include several above and The current question; the contextual intention recognition unit, which performs intention recognition and calculation on the sentence vector of the input sentence, and obtains the intention combined with the context prediction, the integrity of the current question sentence, and the attention weight of the input sentence; Sentence vectors perform intent recognition calculations to obtain the intent of a single sentence; the contextual intent determination unit determines whether the intent of the current question is the intent of a single sentence or the intent of context prediction according to the integrity of the current question sentence and the attention weight of the input sentence. It can effectively identify the importance of the above text to the current question sentence, take into account both the above text and the current question sentence, and ensure the lossless transmission of information.

Description

technical field [0001] The invention belongs to the technical field of computers, and in particular relates to an end-to-end intention recognition system and method based on a memory forgetting device. Background technique [0002] With the development and popularization of the Internet, online shopping is the living habit of most people. People usually need to communicate with merchant customer service to obtain pre-sale, in-sale and after-sale information during shopping. However, traditional real-person customer service cannot satisfy the growing consumer groups. , in customer service, how to provide customers with faster response and more accurate answers is a huge challenge for e-commerce customer service. Robot customer service is the first choice to solve this problem. A response system that provides answers to questions. The semantic recognition system is the core module. Its technical core is natural language processing technology. The earliest method is to identify...

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): G06F40/30G06F16/33G06F16/332
CPCG06F16/3329G06F16/3344
Inventor 张翀江岭
Owner CHENGDU XIAODUO TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap