{"id":"arxiv/2611.24355","url":"https://arxiv.org/abs/2611.24355","tags":["math.NA"],"score":0,"title":"Stable admissible neural classes and convergent physics-informed neural networks","domain":"math.NA","source":"arxiv","authors":["Leevan Ling","Shu Ma","Yizhong Sun"],"reviews":0,"subject":"math.NA","summary":"Residual-based methods, including physics-informed neural networks (PINNs), physics-informed Kolmogorov--Arnold networks, and some physics-informed neural operator architectures, approximate PDE solutions by minimizing a discrete least-squares loss formed from the PDE and boundary residuals over a prescribed trial class. This structure is shared by classical least-squares Kansa collocation, whose kernel trial spaces admit an established convergence theory. For neural trial classes, approximation power alone does not ensure that discrete residuals control the continuous solution error. For…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-11-18","reactions":{}}