Showing 1–2 of 2 results for author: Caramazza, P

A New Paradigm of Reservoir Computing Exploiting Hydrodynamics

Authors: Giulia Marcucci, Piergiorgio Caramazza, Shamit Shrivastava

Abstract: Nonlinear waves have played a historical role in laying the foundations of the science of complexity. Recently, they have also allowed the development of a new reservoir computing paradigm: neuromorphic computing by waves. In these systems, the information transmission acts as the excitation of wave dynamics, whose evolution processes the information to perform complex tasks at low energy consumpt… ▽ More Nonlinear waves have played a historical role in laying the foundations of the science of complexity. Recently, they have also allowed the development of a new reservoir computing paradigm: neuromorphic computing by waves. In these systems, the information transmission acts as the excitation of wave dynamics, whose evolution processes the information to perform complex tasks at low energy consumption. To enable nonlinear hydrodynamic waves to do computing, we designed the Aqua-Photonic-Advantaged Computing Machine by Artificial Neural Networks (Aqua-PACMANN), a system where wave propagation in shallow water is the leading physical phenomenon, and the presence of electronics can be reduced to a CCD camera in detection. We show the realization of an XNOR logic gate as proof of concept of the Aqua-PACMANN's architecture and pave the way to a new class of fluid dynamic neuromorphic computing. △ Less

Submitted 3 February, 2023; originally announced February 2023.

Comments: 10 pages, 4 figures

Neural network identification of people hidden from view with a single-pixel, single-photon detector

Authors: Piergiorgio Caramazza, Alessandro Boccolini, Daniel Buschek, Matthias Hullin, Catherine Higham, Robert Henderson, Roderick Murray-Smith, Daniele Faccio

Abstract: Light scattered from multiple surfaces can be used to retrieve information of hidden environments. However, full three-dimensional retrieval of an object hidden from view by a wall has only been achieved with scanning systems and requires intensive computational processing of the retrieved data. Here we use a non-scanning, single-photon single-pixel detector in combination with an artificial neura… ▽ More Light scattered from multiple surfaces can be used to retrieve information of hidden environments. However, full three-dimensional retrieval of an object hidden from view by a wall has only been achieved with scanning systems and requires intensive computational processing of the retrieved data. Here we use a non-scanning, single-photon single-pixel detector in combination with an artificial neural network: this allows us to locate the position and to also simultaneously provide the actual identity of a hidden person, chosen from a database of people (N=3). Artificial neural networks applied to specific computational imaging problems can therefore enable novel imaging capabilities with hugely simplified hardware and processing times △ Less

Submitted 21 September, 2017; originally announced September 2017.