Abstract:
The exploration and construction of organic chemical reaction networks hold the potential to clarify complex reaction mechanisms. However, this endeavor encounters challenges, particularly in cases requiring computationally intensive quantum mechanical calculations. In recent years, the remarkable advancement of machine learning (ML) technologies has provided a novel approach to address this issue. ML can effectively utilize vast datasets to build and train complex models, enabling prediction and simulation. In this study, we conducted a performance comparison of deep potential for molecular dynamics (DeePMD), recursively embedded atom neural network (REANN), and neural equivariant interatomic potentials (NequIP) based on the Transition1x dataset. The most efficient model, NequIP, was selected for further analysis. Combined with reaction path search methods such as nudged elastic band (NEB) and growing string method (GSM), this model was employed for the identification and exploration of transition states. The results demonstrate that the success ratio of NequIP combined with the NEB method can reach 96.6%, with a mean absolute error (MAE) of 0.32 kcal/mol for barrier prediction. By adding a modest number of data points, NequIP achieves a MAE of 4.01 kcal/mol for barrier prediction in unexplored chemical reaction space. By combining pre-training and fine-tuning model construction strategies, we fine-tuned the models trained with low-precision label data to achieve high precision, significantly reducing the amount of high-precision label data required. In terms of reaction data generation, we improved the YARP method by enumerating reactions of the reactants from the Grambow dataset, the source of the Transition1x dataset, and comparing them with the Transition1x dataset. This effort aims to enhance the reaction database, providing more accurate and reliable data support for machine learning model training, thereby advancing and innovating the field of automated chemical reaction pathway search. In the future, by combining different reaction network search methods, the model can be applied to rapidly explore reaction pathways and construct more comprehensive reaction networks. Machine learning will play a crucial role in accelerating the elucidation of unknown chemical reaction mechanisms and revealing complex reaction networks.
Biography:
Professor Zhu graduated with a Ph.D. from the State Key Laboratory of Precision Spectroscopy in 2013, and was a visiting scholar at the Academia Sinica in Taiwan from 2016 to 2018. His main research direction is the study of the reaction dynamics mechanism of complex chemical systems using computational chemistry and machine learning. In the past five years, he has published more than 80 papers in journals such as Nat. Mach. Intell., Nat. Commun., Nucleic Acids Res., J. Chem. Theory Comput., and J. Chem. Inf. Model.
Seminar Series by the NYU-ECNU Center for Computational Chemistry at NYU Shanghai
This event is open to the NYU Shanghai, NYU, ECNU community and the computational chemistry community.