English

Optimal approximation using complex-valued neural networks

Functional Analysis 2023-10-31 v2 Machine Learning Machine Learning

Abstract

Complex-valued neural networks (CVNNs) have recently shown promising empirical success, for instance for increasing the stability of recurrent neural networks and for improving the performance in tasks with complex-valued inputs, such as in MRI fingerprinting. While the overwhelming success of Deep Learning in the real-valued case is supported by a growing mathematical foundation, such a foundation is still largely lacking in the complex-valued case. We thus analyze the expressivity of CVNNs by studying their approximation properties. Our results yield the first quantitative approximation bounds for CVNNs that apply to a wide class of activation functions including the popular modReLU and complex cardioid activation functions. Precisely, our results apply to any activation function that is smooth but not polyharmonic on some non-empty open set; this is the natural generalization of the class of smooth and non-polynomial activation functions to the complex setting. Our main result shows that the error for the approximation of CkC^k-functions scales as mk/(2n)m^{-k/(2n)} for mm \to \infty where mm is the number of neurons, kk the smoothness of the target function and nn is the (complex) input dimension. Under a natural continuity assumption, we show that this rate is optimal; we further discuss the optimality when dropping this assumption. Moreover, we prove that the problem of approximating CkC^k-functions using continuous approximation methods unavoidably suffers from the curse of dimensionality.

Keywords

Cite

@article{arxiv.2303.16813,
  title  = {Optimal approximation using complex-valued neural networks},
  author = {Paul Geuchen and Felix Voigtlaender},
  journal= {arXiv preprint arXiv:2303.16813},
  year   = {2023}
}

Comments

accepted at NeurIPS 2023

R2 v1 2026-06-28T09:40:14.329Z