$\hat{I}\pm$-SGHN: A Robust Model for Learning Particle Interactions in Lattice Systems

Yixian Gao, Ru Geng, Panayotis Kevrekidis +2

We propose an $\hat{I}\pm$-separable graph Hamiltonian network ($\hat{I}\pm$-SGHN) that reveals complex interaction patterns between particles in lattice systems. Utilizing trajectory data, $\hat{I}\pm$-S…