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

Open answer generation method based on random disturbance network

A random perturbation and openness technology, applied in biological neural network models, computer components, semantic analysis, etc., can solve the problems that openness cannot produce diversification, etc.

Active Publication Date: 2021-02-19
SHANDONG SYNTHESIS ELECTRONICS TECH
View PDF9 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The answer generation model proposed in the existing patents will only generate a standard answer for the same input, and cannot generate diverse answers for open questions

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
  • Open answer generation method based on random disturbance network
  • Open answer generation method based on random disturbance network
  • Open answer generation method based on random disturbance network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0043] This embodiment discloses an open answer generation method based on a random perturbation network, figure 1 The data processing flow chart of the open answer generation method in this method can be summed up in four steps:

[0044] a), using the input device to obtain the original multimodal input data;

[0045] b), serialize the multimodal data, and use the multi-task encoding network to encode the serialized multi-task data;

[0046] c), using a multi-modal perturbation network to correct the original features;

[0047] d) Decode the perturbed features based on the pre-trained language model, and convert the decoding result into text output.

[0048] figure 2 The model structure of this method is shown, including input layer, encoding network, perturbation network, decoding network and output layer. The input layer and the output layer mainly realize the mutual mapping between digital features and text. Except for the input and output layers, the other functional...

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 open answer generation method based on a random disturbance network, and the method comprises the steps: training a disturbance network which generates random disturbance and is integrated with a multi-mode disturbance layer during the training of an answer generation network, correcting the input information in a forward transmission process through the disturbance network so that semantic codes input at any two times are not completely the same, the answer generation network and the disturbance network are cooperatively trained, and the disturbance network does notenable the answer generation model to generate an unreasonable text; based on this, for any two times of input, the method gives answers which are not completely the same but are correct. Diversifiedanswers can be generated at different time, places and situations, the ability is more similar to the performance of human beings, and the model has better environmental adaptability and higher intelligent degree.

Description

technical field [0001] The invention relates to the field of artificial intelligence, in particular to the field of natural language processing of artificial intelligence, in particular to an open answer generation method based on a random perturbation network. Background technique [0002] The text generation model is the core technology in many tasks of natural language processing, such as article summarization, machine translation, article writing, document question answering, open domain dialogue, graphic description, etc. Compared with the generative model, the generative model has higher versatility, and the generative model is the only choice for questions whose answers exceed the scope of the original text of the knowledge base. However, the current text generation model is not mature enough and faces many very serious problems, one of which is that the generated answers are not open. This is because the parameters of the previous models are completely solidified in...

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): G06F40/242G06F16/332G06F40/30G06K9/62G06N3/04
CPCG06F40/242G06F40/30G06F16/3329G06N3/045G06F18/214
Inventor 井焜王太浩张传锋朱锦雷
Owner SHANDONG SYNTHESIS ELECTRONICS TECH
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