{"id":"arxiv/2607.15173","url":"https://arxiv.org/abs/2607.15173","tags":["math.OC"],"score":0,"title":"Landscape analysis for shallow neural networks: Complete classification of critical points for cubic activation and affine target functions","domain":"math.OC","source":"arxiv","authors":["Shokhrukh Ibragimov","Ilkhom Mukhammadiev","Diyora Salimova"],"reviews":0,"subject":"math.OC","summary":"In this paper, we study the optimization landscape induced by the true loss for shallow polynomial neural networks (PNNs) with $\\mathfrak{h} \\in \\mathbb{N}$ neurons on the hidden layer, one-dimensional input and output layers, and a monomial activation of degree $d \\in \\mathbb{N}$, trained against a non-constant affine linear target function. Our first main result provides for arbitrary activation degree $d$ a sharp existence/non-existence criterion for \\emph{global minimizers} with necessary structural conditions. We show that the infimum of the loss is always zero and achievable with at…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-16","reactions":{}}