{"id":"arxiv/2606.06046","url":"https://arxiv.org/abs/2606.06046","tags":["math.NA","cs.LG"],"score":0,"title":"Learning solution operators of PDEs with sparse approximation methods","domain":"math.NA","source":"arxiv","authors":["Sebastian Neumayer","Daniel Potts","Fabian Taubert"],"reviews":0,"subject":"math.NA","summary":"We investigate the approximation of solution operators for partial differential equations (PDEs) using sparse high-dimensional techniques. Building on a dimension-incremental framework, we combine product basis expansions with sparse recovery methods, specifically orthogonal matching pursuit (OMP), to substantially reduce the required sample size compared with a previously considered cubature-based approach. We evaluate the resulting method numerically on several examples, comparing it against both cubature-based sparse approximation and Fourier neural operators in terms of accuracy,…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-06-04","reactions":{}}