{"id":"arxiv/2601.08709","url":"https://arxiv.org/abs/2601.08709","tags":["math.NA","cs.LG"],"score":0,"title":"Multi-Preconditioned LBFGS for Training Finite-Basis PINNs","domain":"math.NA","source":"arxiv","authors":["Marc Salvadó-Benasco","Aymane Kssim","Alexander Heinlein","Rolf Krause","Serge Gratton","Alena Kopaničáková"],"reviews":0,"subject":"math.NA","summary":"A multi-preconditioned LBFGS (MP-LBFGS) algorithm is introduced for training finite-basis physics-informed neural networks (FBPINNs). The algorithm is motivated by the nonlinear additive Schwarz method and exploits the domain-decomposition-inspired additive architecture of FBPINNs, in which local neural networks are defined on subdomains, thereby localizing the network representation. Parallel, subdomain-local quasi-Newton corrections are then constructed on the corresponding local parts of the architecture. A key feature is a novel nonlinear multi-preconditioning mechanism, in which…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-01-13","reactions":{}}