CrystAF is introduced, an all-atom crystal flow-map generation model, and CrystAF with the UMA interatomic potential is used to systematically study where physics should enter, suggesting a simple principle: learn reusable physical alignment into the generator, and reserve inference-time physics for residual constraint...
This work proposes Visual Latent Structural Reasoning (VLSR), an end-to-end framework that jointly learns localization and reasoning from molecular images, and central to the approach is a localize-then-reason strategy.
A novel, large-scale reasoning dataset of reaction mechanisms, and the FukuyamaBench, a difficult benchmark derived from Fukuyama's Advanced Organic Reaction Mechanism book, to rigorously evaluate model performance on hierarchical mechanism reasoning, demonstrate that mechanism-aware training substantially enhances che...
Xingyu Dang, Haocheng Tang, Junmei Wang et al.· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.