Jul 2026
Relation-Aware Pretraining and Reaction Center Modeling for Chemical Reaction Graph Representation Learning.
Evaluations across three downstream tasks show that KAGT achieves strong performance relative to existing baselines in reaction classification, reaction condition prediction, and yield prediction, supporting KAGT as a transferable representation framework for AI-driven chemical synthesis.
Jian-Bo Qiao, Ke-Fei Li, Jun-Ru Jin et al.
· Journal of Chemical Theory a... · 1 citation