{"id":"arxiv/2610.20825","url":"https://arxiv.org/abs/2610.20825","tags":["math.AG","cs.LG"],"score":0,"title":"Linear Independence of Polynomial Compositions with Bounded Input Degree","domain":"math.AG","source":"arxiv","authors":["Xiaotian Chang","Yacoub Hendi"],"reviews":0,"subject":"math.AG","summary":"We prove the weak composition-independence conjectures of Kohn, Marchetti, Massarenti and Mella. The proof uses deformations of rational curves and a single kernel of the linearized equation. Riemann-Roch gives a contradiction from tangent fields with controlled poles. By settling these conjectures, we prove that a neural network with a general polynomial activation function of sufficiently large degree is identifiable.","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-14","reactions":{}}