English

Optimal Non-Asymptotic Edgeworth Expansions for Multivariate Neural Network Outputs

Machine Learning 2026-05-26 v1 Machine Learning Probability

Abstract

Finite-width fully connected neural networks with Gaussian-initialized weights deviate from their infinite-width Gaussian limit, exhibiting non-vanishing higher-order cumulants. We approximate these deviations, for a neural network evaluated in a finite number of inputs, using multidimensional Edgeworth expansions of arbitrary order 4m14m-1, with mNm\in\mathbb{N}. Assuming that the corresponding Gaussian limit has an invertible covariance matrix and that the activation function is polynomially bounded, we establish a bound of order nmn^{-m} on the total variation distance between the law of the true network output and its Edgeworth approximation, with matching lower bounds. As an application, we quantify the error in Bayesian posterior distributions when the prior is replaced by its Edgeworth expansion. Our results are more general and also apply to sequences of conditionally Gaussian vectors converging to a Gaussian vector with invertible covariance.

Keywords

Cite

@article{arxiv.2605.24072,
  title  = {Optimal Non-Asymptotic Edgeworth Expansions for Multivariate Neural Network Outputs},
  author = {Lucia Celli},
  journal= {arXiv preprint arXiv:2605.24072},
  year   = {2026}
}

Comments

34 pages, 2 figures