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

Molecule attribute prediction method based on artificial neural network

A technology of artificial neural network and prediction method, applied in biological neural network model, chemical property prediction, molecular design, etc., can solve problems such as inability to use data, and achieve high precision and high efficiency.

Active Publication Date: 2019-03-12
UNIV OF SCI & TECH OF CHINA
View PDF5 Cites 16 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

At the same time, this research field has accumulated a lot of relevant data, but most of the methods cannot use these existing data

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
  • Molecule attribute prediction method based on artificial neural network
  • Molecule attribute prediction method based on artificial neural network
  • Molecule attribute prediction method based on artificial neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0082] Take the U of the following three molecules 0 Forecasting as an example, they all come from the commonly used QM9 data set in the world, and the unit is eV. We use the QM9 data set as the training set, the training method is as above, and then use the model obtained after training to predict the following molecules. The standard to measure the error is the absolute error, which is the absolute value of the difference between the predicted value and the actual value.

[0083] (1)CH 4 The true value of is -17.1717476eV, the predicted value is -17.1681695eV, and the error is 0.0035781eV. (2) NH 3 The true value of is -12.0055513eV, the predicted value is -12.0187658eV, and the error is 0.0132145eV.

[0084] (3) The actual value of HOH is -9.2401279eV, the predicted value is -9.2371538eV, and the error is 0.0029741eV.

[0085] And the average error of each attribute prediction of this method on the entire QM9 data set is given in the table below.

[0086]

[0087] ...

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 provides a molecule attribute prediction method based on an artificial neural network. The molecule attribute prediction method based on the artificial neural network comprises the following steps: S1) preprocessing of molecule data: acquiring atomic space representation and atomic composition representation through a method for data structure representation of graphs; S2) model establishment: acquiring representation of different stages of molecules from the atomic space representation and the atomic composition representation through a multilayered convolutional neural network,and combining the representations of the different stages of the molecules to obtain a model; and S3) predicting the molecule attribute according to the model. Compared with the prior art, the molecule attribute prediction method based on the artificial neural network has the advantages that the multistage convolutional neural network is used, the relation between the molecule attribute and spacecomposition can be learnt from information of known data and the multistage structure of the molecules, and is used for predicting the related attributes of unknown molecules, and therefore, the speed and the precision are good.

Description

technical field [0001] The invention belongs to the technical field of materials science, and in particular relates to a method for predicting molecular properties based on an artificial neural network. Background technique [0002] From drug development to material development, molecular discovery is inseparable. In order to find molecules with specific properties to meet the needs of applications, the general method is to traverse an unknown set of possible molecules (called chemical space), during the traversal process, researchers use various methods to predict molecular properties properties, and if a molecule is found to meet the requirements, it is recorded for further study. For example, predictions of the energy properties of molecules can help researchers find stable molecules. [0003] However, the chemical space is often very large, and a widely used chemical space has more than 160 billion molecules. Therefore, a rapid method for determining molecular propert...

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
IPC IPC(8): G16C20/30G16C20/40G16C20/50G06N3/04
CPCG06N3/045
Inventor 刘淇陈恩红陆承镪王超黄振亚
Owner UNIV OF SCI & TECH OF CHINA
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