介绍:
播客系列标题：《AI化学前沿：分子革命》
单集标题：第三集：AI预测魔法：选择最佳化学配方
Episode 3: Predicting Reaction Yield: Towards Data-Driven Optimization(Total Time: ~35 minutes)
Guests:
Dr. Abigail Doyle(Princeton University) - (Continuing from Episode 1, focusing on yield prediction aspects)
Dr. Paul Schwaller(IBM Research) - Expert in reaction yield prediction using deep learning and NLP.
Focus:In-depth discussion on machine learning models for yield prediction (Section 5). Explore different ML algorithms used (RF, SVM, Naive Bayes, ANNs). Discuss case studies like Buchwald-Hartwig amination [71], tungsten-catalyzed epoxidation [72], Suzuki-Miyaura coupling [74], and the use of BERT encoders [73]. Discuss the importance of data size and quality for prediction accuracy.
大家还在听
