首页|Shanghai Jiao Tong University Reports Findings in Machine Learning (Prediction o f chemical reaction yields with large-scale multiview pre-training)
Shanghai Jiao Tong University Reports Findings in Machine Learning (Prediction o f chemical reaction yields with large-scale multiview pre-training)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-New research on Machine Learning is th e subject of a report. According to news reporting originating from Shanghai, Pe ople's Republic of China, by NewsRx correspondents, research stated, "Developing machine learning models with high generalization capability for predicting chem ical reaction yields is of significant interest and importance. The efficacy of such models depends heavily on the representation of chemical reactions, which h as commonly been learned from SMILES or graphs of molecules using deep neural ne tworks." Financial support for this research came from National Natural Science Foundatio n of China. Our news editors obtained a quote from the research from Shanghai Jiao Tong Univ ersity, "However, the progression of chemical reactions is inherently determined by the molecular 3D geometric properties, which have been recently highlighted as crucial features in accurately predicting molecular properties and chemical r eactions. Additionally, large-scale pre-training has been shown to be essential in enhancing the generalization capability of complex deep learning models. Base d on these considerations, we propose the Reaction Multi-View Pre-training (ReaM VP) framework, which leverages self-supervised learning techniques and a two-sta ge pre-training strategy to predict chemical reaction yields. By incorporating m ulti-view learning with 3D geometric information, ReaMVP achieves state-of-the-a rt performance on two benchmark datasets. Notably, the experimental results indi cate that ReaMVP has a significant advantage in predicting out-of-sample data, s uggesting an enhanced generalization ability to predict new reactions. Scientifi c Contribution: This study presents the ReaMVP framework, which improves the gen eralization capability of machine learning models for predicting chemical reacti on yields. By integrating sequential and geometric views and leveraging self-sup ervised learning techniques with a two-stage pre-training strategy, ReaMVP achie ves state-of-the-art performance on benchmark datasets."
ShanghaiPeople's Republic of ChinaAs iaCyborgsEmerging TechnologiesMachine LearningSupervised Learning
NETL
NSTL
