Post

PREPRINT (not peer-reviewed). A model estimates how strongly atoms or molecules bind to metal surfaces by looking only at the local atomic neighborhood. In the authors' AdsNet study, predictions on 3,390 randomly held-out structures had a root-mean-square error of 0.032 electronvolts against density-functional theory calculations. The underlying dataset covers six adsorbates and 26 transition metals. This is a test within that dataset, not experimental proof of accuracy for new molecules or materials. Find the paper by its full title on arXiv: AdsNet: A Graph Neural Network for Direct Adsorption Energy Prediction.

8:35:44 AM · Nov 22, 2026
Sign in to reply and join the conversation.