Methods, systems, devices, and media for optimization of chemical reaction yields

By constructing a similarity-constrained representation mapping model and a substructure masking method, the contribution of substructures in chemical reactions is quantified, solving the problem of inefficient chemical reaction yield optimization in existing technologies, and realizing efficient yield improvement and interpretable optimization strategies.

CN121958905BActive Publication Date: 2026-06-09UNIV OF SCI & TECH OF CHINA
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
UNIV OF SCI & TECH OF CHINA
Filing Date
2026-04-02
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing technologies rely on researchers' experience to optimize chemical reaction yields, resulting in low efficiency, difficulty in understanding the chemical mechanisms of reaction systems, and a lack of explanatory power in black-box models, making it impossible to effectively translate them into targeted modification strategies.

Method used

A similarity-constrained representation mapping model is constructed to map molecular fingerprints into reaction latent representations rich in reaction yield semantic information. The contribution of each substructure to the predicted yield is quantified by the substructure masking method, thereby achieving a white-box interpretation of the structure-performance relationship.

Benefits of technology

Precisely identifying specific substructures that promote or inhibit reaction yields can improve chemical reaction yields, shorten the cycle from laboratory to industrial application, and increase R&D efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121958905B_ABST
    Figure CN121958905B_ABST
Patent Text Reader

Abstract

The present application relates to the field of artificial intelligence, in particular to a chemical reaction yield optimization method, system, device and medium, comprising: obtaining training samples and yield labels of a target chemical reaction; generating molecular fingerprints of each training sample as initial representation, and determining substructures contained in the training samples; training a to-be-trained representation mapping model supervised by the yield labels, so that the similarity of the mapped features is consistent with the yield similarity, and obtaining a representation mapping model; mapping the initial representation to a reaction hidden representation through the representation mapping model; inputting the reaction hidden representation into a yield prediction model to obtain a predicted yield; calculating the contribution degree of each substructure by masking the influence of the corresponding code of the substructure on the predicted yield; and optimizing the target chemical reaction according to the contribution degree. The present application can accurately locate specific substructures that promote or inhibit the yield, and introduce beneficial substructures or avoid adverse substructures, thereby improving the yield of the target chemical reaction.
Need to check novelty before this filing date? Find Prior Art