Chemical reaction conversion rate prediction method and system based on deep learning, and medium

Abstract

The invention discloses a chemical reaction conversion rate prediction method and system based on deep learning and a medium, and the method comprises the steps: selecting a reactant type A which has the greatest influence on the chemical reaction conversion rate from the types of reactants participating in a chemical reaction, carrying out word segmentation and feature extraction on reactants of the reactant type A and a chemical reaction formula R corresponding to the chemical reaction, and then solving deep characterization information Ta of the reactants of the reactant type A, deep characterization information Tr of the chemical reaction formula R corresponding to the chemical reaction and relation information Tm of the reactants and the reaction formulas through an attention mechanism; and finally, fusing the deep characterization information Ta of the reactant of the reactant type A, the deep characterization information Tr of the chemical reaction formula R corresponding to the chemical reaction and the relation information Tm of the reactant and the reaction formula by using a wide learning system, and predicting the conversion rate of the chemical reaction. The prediction precision of the chemical reaction conversion rate is greatly improved.

Description

technical field [0001] The present invention relates to the technical field of chemical reactions, in particular to methods, systems and media for predicting conversion rates of chemical reactions based on deep learning. Background technique [0002] The reaction conversion rate is the ratio of the actual reaction product of a chemical reaction to the theoretical reaction product. Under ideal conditions, the conversion rate of a chemical reaction should be 100%. However, under realistic conditions, affected by conditions such as temperature and concentration, most reactions conversion rate is less than 100%. [0003] In the design of a chemical reaction process, it is common to generate the desired product through several steps or even dozens of steps. No matter which step in the reaction process the conversion rate is too low, it will have a fatal impact on the entire reaction process due to the cumulative effect. Therefore, it is particularly important to design chemical...

AI Technical Summary

Problems solved by technology

[0004] In recent years, deep learning has achieved good results in various fields. Many people have begun to try to use deep learning methods to predict reaction conversion rates to help design the entire chemical reaction process. However, the existing chemical reaction conversion rate prediction models Most of them are for a specific reaction type, such as [1] Ahneman D T, Estrada J G, Lin S, Dreher S D and Doyle A G 2018 Predicting reaction performance in C–Ncross-coupling using machine learning Science 360 ​​186–90. And [2] Chuang K Vand KeiserM J 2018 Comment on "Predicting reaction performance in C–N cross-coupling using machine learning" Science 362 6416. This type of model is not very versatile

Recently, although a general conversion rate prediction model based on natural language processing technology [3] Schwaller P, Vaucher A C, Laino T, et al. Prediction of chemical reaction yields using deep learning [J]. Machine Learning: Science and Technology, 2021,2 (1):015016 was also proposed, but this method directly takes the entire reaction formula as the input of the model, does not fully consider the influence of different reactants on the reaction formula and conversion rate, and does not make full use of the reactant information

Images