Regio-selectivity prediction with a machine-learned reaction representation and on-the-fly quantum mechanical descriptors

Abstract

Integrating feature learning and on-the-fly feather engineering enables fast and accurate reacitvity predictions using large or small dataset.

DeepReac+: deep active learning for quantitative modeling of organic chemical reactions

“…random forest, have an advantage in the low data regime. As shown by a previous report 95 , it is likely that GNNs can be further enhanced by relevant descriptors. Thus, inclusion of quantum chemical descriptors into DeepReac+ may be a future solution.…”

Real-time prediction of 1H and 13C chemical shifts with DFT accuracy using a 3D graph neural network

“…These GNN-derived atomic descriptors impose low computational cost such that we anticipate future utility in related prediction tasks of organic reactivity and selectivity, for example in combination with other machine-learned representations. 87 …”

Computational Organometallic Catalysis: Where We Are, Where We Are Going

“…[60] A reactivity descriptor database based on QM calculations of 130,000 organic molecules was used to predict regio-selectivity for three general types of substitution reactions. [61] Trends surrounding the thermodynamics of the hydroformylation reaction catalyzed by group 9 metals bearing phosphine ligands have been analyzed using a data-driven inspired workflow (data-powered volcano plots). The total data set used consisted of 1510 catalytic cycles derived from DFT computations and 491 catalytic cycles derived from machine learned profiles.…”

“…Different ML algorithms were used for different purposes, including neural networks for accuracy, Gaussian processes for transferability, and Gradient boosting for explainability [60] . A reactivity descriptor database based on QM calculations of 130,000 organic molecules was used to predict regio‐selectivity for three general types of substitution reactions [61] . Trends surrounding the thermodynamics of the hydroformylation reaction catalyzed by group 9 metals bearing phosphine ligands have been analyzed using a data‐driven inspired workflow (data‐powered volcano plots).…”