{"id":"arxiv/2602.09959","url":"https://arxiv.org/abs/2602.09959","tags":["math.ST","cs.LG","stat.ML"],"score":0,"title":"Statistical-Computational Trade-offs in Learning Multi-Index Models via Harmonic Analysis","domain":"math.ST","source":"arxiv","authors":["Hugo Latourelle-Vigeant","Theodor Misiakiewicz"],"reviews":0,"subject":"math.ST","summary":"We study the problem of learning multi-index models (MIMs), where the label depends on the input $\\boldsymbol{x} \\in \\mathbb{R}^d$ only through an unknown $\\mathsf{s}$-dimensional projection $\\boldsymbol{W}_*^\\mathsf{T} \\boldsymbol{x} \\in \\mathbb{R}^\\mathsf{s}$. Exploiting the equivariance of this problem under the orthogonal group $\\mathcal{O}_d$, we obtain a sharp harmonic-analytic characterization of the learning complexity for MIMs with spherically symmetric inputs -- which refines and generalizes previous Gaussian-specific analyses. Specifically, we derive statistical and computational…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-02-10","reactions":{}}