Journal Article

Noise-aware training of neuromorphic dynamic device networks

Manneschi, Luca, Donskikh, Denis, Stepney, Susan, Swindells, Charles, Gui, Lai, Ellis, Matthew O. A., Vidamour, Ian T., Venkat, Guru, Branford, Will R., Hayward, Thomas J., Vasilaki, Eleni, Sonawala, Daanish, Griffin, David, Stenning, Kilian D., Gartside, Jack C., Hariga, Dana, Donati, Elisa

2025

Overview

In materio computing offers the potential for widespread embodied intelligence by leveraging the intrinsic dynamics of complex systems for efficient sensing, processing, and interaction. While individual devices offer basic data processing capabilities, networks of interconnected devices can perform more complex and varied tasks. However, designing such networks for dynamic tasks is challenging in the absence of physical models and accurate characterization of device noise. We introduce the Noise-Aware Dynamic Optimization (NADO) framework for training networks of dynamical devices, using Neural Stochastic Differential Equations (Neural-SDEs) as differentiable digital twins to capture both the dynamics and stochasticity of devices with intrinsic memory. Our approach combines backpropagation through time with cascade learning, enabling effective exploitation of the temporal properties of physical devices. We validate this method on networks of spintronic devices across both temporal classification and regression tasks. By decoupling device model training from network connectivity optimization, our framework reduces data requirements and enables robust, gradient-based programming of dynamical devices without requiring analytical descriptions of their behaviour. Dynamic systems show promise for physical neural networks, but gradient based optimization requires mathematical models. Here, the authors present a data-driven framework for optimizing networks of arbitrary dynamic systems which is robust to noise, and enables tasks such as neuroprosthetic control.

• Publisher: Nature Publishing Group UK,Nature Publishing Group,Nature Portfolio
• Subject: 639/705/117 / 639/766/119/1001 / Algorithms / Back propagation / Back propagation networks / Complex systems / Data processing / Decoupling / Differential equations / Digital computers / Digital twins / Dynamical systems / Humanities and Social Sciences / Mathematical models / Memory devices / multidisciplinary / Neural networks / Neural prostheses / Optimization / Prosthetics / Robustness / Science / Science (multidisciplinary) / Simulation / Stochasticity / Task complexity