In the Research Article ( e18560 ), Xiaohua Liu, Xiaoming Feng, Xin Hong, and co-workers report an integrated data platform for N,N' -dioxide/metal-catalyzed asymmetric Michael additions. Featuring over 2,000 curated reactions, it combines statistical analysis with mechanistically informed machine learning. By integrating intermediate-based augmentation and similarity-weighted tuning, the model achieves accurate selectivity prediction, validated by the successful design of a new enantioselective transformation.
Tang et al. (Tue,) studied this question.