Matching of signal, noise and hardware timescales for filtering and forecasting of correlated noise signals
Abstract
Physical reservoir computing exploits the nonlinear dynamics of physical systems to process time-dependent data with greater energy efficiency than conventional machine learning approaches. However, physical reservoirs have fixed intrinsic response timescales, whereas real-world signals combine deterministic and stochastic components across multiple timescales. Here we show, using a nanoporous niobium oxide reservoir, synthetic noisy signals and cryptocurrency-price volatility, that the relationship among noise correlation time, reservoir memory and forecast horizon determines whether correlated noise is filtered or predicted. Noise varying faster than the relevant reservoir memory and forecast horizon is averaged by the reservoir, whereas the temporal structure of slower-varying noise is sufficient for algorithmic forecasting. We introduce the reservoir memory horizon and forecasting regime index to distinguish these operating regimes. These contributions demonstrate that timescale matching can guide the encoding of input time series and development of physical reservoir architectures that filter, analyse and predict stochastic signal components across distinct temporal scales.
Related
- Hardware-Aware Design and Layout Selection in Quantum Reservoir Computing for Insurance Claims Forecasting
- Perspective on Magnetic Nanoparticle Modeling: Interactions, Timescales and Regimes
- Dependence of quantum timescales on symmetry
- Building a Better Beta: Nucleation and Timescales in Cosmological Phase Transitions
- Spectral-Gap Bounds and Timescales for Purity Loss in Hamiltonian--Pointer Interactions
BibTeX
@misc{donald2026matching,
title = {{Matching of signal, noise and hardware timescales for filtering and forecasting of correlated noise signals}},
author = {Joshua Donald and Alex Gabbitas and Arthur G. T. Coveney and Sergey Savel'ev and Pavel Borisov},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.10037},
eprint = {2610.10037},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.10037},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Matching of signal, noise and hardware timescales for filtering and forecasting of correlated noise signals
AU - Joshua Donald
AU - Alex Gabbitas
AU - Arthur G. T. Coveney
AU - Sergey Savel'ev
AU - Pavel Borisov
PY - 2026
DA - 2026/10/07
PB - arXiv
DO - 10.48550/arxiv.2610.10037
UR - https://arxiv.org/abs/2610.10037
AB - Physical reservoir computing exploits the nonlinear dynamics of physical systems to process time-dependent data with greater energy efficiency than conventional machine learning approaches. However, physical reservoirs have fixed intrinsic response timescales, whereas real-world signals combine deterministic and stochastic components across multiple timescales. Here we show, using a nanoporous niobium oxide reservoir, synthetic noisy signals and cryptocurrency-price volatility, that the relationship among noise correlation time, reservoir memory and forecast horizon determines whether correlated noise is filtered or predicted. Noise varying faster than the relevant reservoir memory and forecast horizon is averaged by the reservoir, whereas the temporal structure of slower-varying noise is sufficient for algorithmic forecasting. We introduce the reservoir memory horizon and forecasting regime index to distinguish these operating regimes. These contributions demonstrate that timescale matching can guide the encoding of input time series and development of physical reservoir architectures that filter, analyse and predict stochastic signal components across distinct temporal scales.
ER -