English
 
Help Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT

Released

Paper

Fully Non-Linear Neuromorphic Computing with Linear Wave Scattering

MPS-Authors

Wanjura,  Clara C.
Marquardt Division, Max Planck Institute for the Science of Light, Max Planck Society;

/persons/resource/persons201125

Marquardt,  Florian
Marquardt Division, Max Planck Institute for the Science of Light, Max Planck Society;

External Resource
No external resources are shared
Fulltext (restricted access)
There are currently no full texts shared for your IP range.
Fulltext (public)

2308.16181.pdf
(Any fulltext), 849KB

Supplementary Material (public)

Bildschirmfoto 2023-09-04 um 11.18.40.png
(Supplementary material), 45KB

Citation

Wanjura, C. C., & Marquardt, F. (2023). Fully Non-Linear Neuromorphic Computing with Linear Wave Scattering. arXiv, 2308.16181.


Cite as: https://hdl.handle.net/21.11116/0000-000D-A7E6-3
Abstract
The increasing complexity of neural networks and the energy consumption associated with training and inference create a need for alternative neuromorphic approaches, e.g. using optics. Current proposals and implementations rely on physical non-linearities or opto-electronic conversion to realise the required non-linear activation function. However, there are significant challenges with these approaches related to power levels, control, energy-efficiency, and delays. Here, we present a scheme for a neuromorphic system that relies on linear wave scattering and yet achieves non-linear processing with a high expressivity. The key idea is to inject the input via physical parameters that affect the scattering processes. Moreover, we show that gradients needed for training can be directly measured in scattering experiments. We predict classification accuracies on par with results obtained by standard artificial neural networks. Our proposal can be readily implemented with existing state-of-the-art, scalable platforms, e.g. in optics, microwave and electrical circuits, and we propose an integrated-photonics implementation based on racetrack resonators that achieves high connectivity with a minimal number of waveguide crossings.