[We-P] 2D materials

[We-P-48] Interpretable Physics-Constrained Machine Learning for Identifying Multi-Nanobubble Configurations in Graphene

〇Jihye Kim¹, Nojoon Myoung², Taegeun Song¹ (1. Kongju Natl. Univ. (Korea), 2. Chosun Univ. (Korea))

Physics-constrained ML identifies multiple graphene nanobubbles from composite DOS spectra by embedding an additive physical prior into a neural decomposition model. The method is accurate, interpretable, and efficient for graphene nanodevice characterization.