Reasoning model-oriented object-relationship construction method and device

An object relationship and construction method technology, applied in the computer field, can solve problems such as poor applicability and complex deep learning model conditions, and achieve the effect of improving the efficiency of joint training

Inactive Publication Date: 2021-08-06
POTEVIO INFORMATION TECH CO LTD
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In other words, the conditions for the establishment of the deep learning model are more complicated and the applicability is poor

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
  • Reasoning model-oriented object-relationship construction method and device
  • Reasoning model-oriented object-relationship construction method and device
  • Reasoning model-oriented object-relationship construction method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment approach

[0027] In some implementation manners, there may be many ways to combine target objects corresponding to different types of data sources in step S102. It can be understood that the types of data sources mixed in the mixed samples may be different in the way of combining the target objects. The following describes an optional implementation of target object combination when the mixed sample contains image data and natural language data, including:

[0028] S2011. Combining all the image target objects corresponding to the image data in pairs to obtain several image secondary combination pairs;

[0029] S2012. Combining all the natural language target objects corresponding to the natural language data in pairs to obtain several natural language secondary combination pairs;

[0030] S2013. Randomly combine the image secondary combination pair with the natural language secondary combination pair to obtain a combination pair.

[0031] For details, please refer to image 3 , go t...

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

Embodiments of the present invention provide a reasoning model-oriented object relationship construction method and device. In the method, objects are extracted for different data sources based on different reasoning models to obtain corresponding target objects, and then various types of data are correspondingly The target objects are combined according to the preset rules, and the obtained combination pairs are used as the input data to be trained, thus solving the problem of object relationship construction of multi-type mixed source data, and providing a simple, effective and easy-to-implement method for the training of inference models. The construction method of the complex object relationship is conducive to improving the joint training efficiency of data targets from different data sources.

Description

technical field [0001] The embodiments of the present invention relate to the field of computer technology, and in particular to a reasoning model-oriented object relationship construction method and device. Background technique [0002] A trained neural network can apply what it has learned to tasks in the digital world in general—recognizing images, recognizing speech, detecting disease, or recommending advertisements, to name a few. This faster and more efficient way for a neural network to make inferences about new data it acquires based on what it was trained on is known as inference. Inference can be generated without training, and inference tasks often do not require all the infrastructure of their training schemes to achieve good results. The goal of training (similar to human receiving education) is knowledge acquisition, and the training of neural network is quite different from the process of human receiving education. Neurons in the human brain can be connected...

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): G06N3/08G06N5/04
CPCG06N3/082G06N5/04
Inventor 李乃鹏
Owner POTEVIO INFORMATION 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