{"id":"arxiv/2605.15881","url":"https://arxiv.org/abs/2605.15881","tags":["math.DS","cs.AI","physics.comp-ph"],"score":0,"title":"Symplectic Neural Operators for Learning Infinite Dimensional Hamiltonian Systems","domain":"math.DS","source":"arxiv","authors":["Yeang Makara","Yusuke Tanaka","Takashi Matsubara","Takaharu Yaguchi"],"reviews":0,"subject":"math.DS","summary":"The modeling and simulation of infinite-dimensional Hamiltonian systems are central problems in mathematical physics and engineering, however they pose significant computational and structural challenges for standard data-driven architectures. In this work, we introduce the Symplectic Neural Operator, a neural operator architecture designed to preserve the symplectic structure intrinsic to Hamiltonian PDEs. We provide a theoretical characterization of their symplecticity and establish a rigorous long-term stability result based on the combination of symplectic structure preservation and…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-05-15","reactions":{}}