Modeling neural networks with active optical devices

Alastair D. McAulay

doi:10.1117/12.619354

15 September 2005 Modeling neural networks with active optical devices

Alastair D. McAulay

Proceedings Volume 5908, Optical Information Systems III; 59080G (2005) https://doi.org/10.1117/12.619354
Event: Optics and Photonics 2005, 2005, San Diego, California, United States

Abstract

Computers perform poorly in comparison with the human brain for everyday tasks because they evolved from a 1940's number crunching architecture. New architectures are needed, more appropriate for current everyday tasks: web database search, network communication and recognition. For a new computer architecture, we consider the Wilson Cowan biological neural network (BNN) model of the brain which couples excitatory and inhibitory neurons in an additive neural model with sigmoid nonlinearity. We analyze the Wilson Cowan equations to demonstrate supercritical Hopf bifurcation that switches the neural oscillator between off and a stable oscillation---a basic operation in the brain. To mimic the Wilson Cowan neuron, we propose a novel all-optical artificial neural network (ANN) architecture in which semiconductor optical amplifiers (SOAs) perform addition and provide optical injection to a laser diode. We analyze the proposed all-optical system to show that it also performs supercritical Hopf bifurcation, although in this case with 3-D time trajectories rather than 2-D ones. We conclude that the all-optical system has as much dynamic capability as the mathematical brain model. This research aims to stimulate the development of physical analogs to the brain for future computing and better understanding of the brain.

Citation Download Citation

Alastair D. McAulay "Modeling neural networks with active optical devices", Proc. SPIE 5908, Optical Information Systems III, 59080G (15 September 2005); https://doi.org/10.1117/12.619354

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available