AURA: Angular Update Rate Adaptation for training complex-valued neural networks
Abstract
Complex-valued neural networks (CVNNs) are increasingly adopted for complex-valued data; however, they are often trained with first-order optimizers inherited from the real-valued case. The efficiency of these methods depends largely on the step size, and their step-size rules ignore the angular information available in the complex plane. We address step-size adaptation in the complex domain by introducing AURA (Angular Update Rate Adaptation), a per-parameter step-size adaptation that can be added on top of any first-order optimizer, and removed from it, without altering its update direction. AURA measures the agreement between consecutive updates of each complex parameter, in length, alignment, and sense of rotation, and enlarges the step when they are consistent and reduces it when they are not. It requires no additional gradient evaluations and only inexpensive vector operations per step. We combine AURA with Adam and Muon and compare the resulting methods with well-known first-order optimizers on four test cases of increasing complexity, ranging from the approximation of scalar complex functions to physics-informed training. Fully connected neural networks are used throughout this work. All hyperparameters other than the step size are held fixed across test cases; for one case, we also tune the hyperparameters of each optimizer under the same budget. Our empirical tests show that AURA improves the convergence of its base optimizer in most cases with a small per-step overhead, and we identify the conditions under which it fails to do so.
Related
- Lazy training of quantum physics informed neural networks
- Subspace Levenberg Marquardt Algorithms in Training Neural Networks
- Two-level domain-decomposition AdaGrad method for scalable training of graph neural networks
- Width and Depth Limits Commute During Training in Linear Residual Networks
- Learning Changes the World and Then What Was Learned has Changed: Training and Deploying Neural Networks in the Control Loop
BibTeX
@misc{ballini2026aura,
title = {{AURA: Angular Update Rate Adaptation for training complex-valued neural networks}},
author = {Enrico Ballini and Allan Peter Engsig-Karup and Tito Andriollo},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2609.25914},
eprint = {2609.25914},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.25914},
}
RIS
TY - UNPB
TI - AURA: Angular Update Rate Adaptation for training complex-valued neural networks
AU - Enrico Ballini
AU - Allan Peter Engsig-Karup
AU - Tito Andriollo
PY - 2026
DA - 2026/10/20
PB - arXiv
DO - 10.48550/arxiv.2609.25914
UR - https://arxiv.org/abs/2609.25914
AB - Complex-valued neural networks (CVNNs) are increasingly adopted for complex-valued data; however, they are often trained with first-order optimizers inherited from the real-valued case. The efficiency of these methods depends largely on the step size, and their step-size rules ignore the angular information available in the complex plane. We address step-size adaptation in the complex domain by introducing AURA (Angular Update Rate Adaptation), a per-parameter step-size adaptation that can be added on top of any first-order optimizer, and removed from it, without altering its update direction. AURA measures the agreement between consecutive updates of each complex parameter, in length, alignment, and sense of rotation, and enlarges the step when they are consistent and reduces it when they are not. It requires no additional gradient evaluations and only inexpensive vector operations per step. We combine AURA with Adam and Muon and compare the resulting methods with well-known first-order optimizers on four test cases of increasing complexity, ranging from the approximation of scalar complex functions to physics-informed training. Fully connected neural networks are used throughout this work. All hyperparameters other than the step size are held fixed across test cases; for one case, we also tune the hyperparameters of each optimizer under the same budget. Our empirical tests show that AURA improves the convergence of its base optimizer in most cases with a small per-step overhead, and we identify the conditions under which it fails to do so.
ER -