Photonic systems for high-performance information processing have attracted
renewed interest. Neuromorphic silicon photonics has the potential to integrate
processing functions that vastly exceed the capabilities of electronics. We
report first observations of a recurrent silicon photonic neural network, in
which connections are configured by microring weight banks. A mathematical
isomorphism between the silicon photonic circuit and a continuous neural
network model is demonstrated through dynamical bifurcation analysis.
Exploiting this isomorphism, a simulated 24-node silicon photonic neural
network is programmed using "neural compiler" to solve a differential system
emulation task. A 294-fold acceleration against a conventional benchmark is
predicted. We also propose and derive power consumption analysis for
modulator-class neurons that, as opposed to laser-class neurons, are compatible
with silicon photonic platforms. At increased scale, Neuromorphic silicon
photonics could access new regimes of ultrafast information processing for
radio, control, and scientific computing.
%0 Generic
%1 tait2016neuromorphic
%A Tait, Alexander N.
%A de Lima, Thomas Ferreira
%A Zhou, Ellen
%A Wu, Allie X.
%A Nahmias, Mitchell A.
%A Shastri, Bhavin J.
%A Prucnal, Paul R.
%D 2016
%K quantumnuralnet
%R 10.1038/s41598-017-07754-z
%T Neuromorphic Silicon Photonic Networks
%U http://arxiv.org/abs/1611.02272
%X Photonic systems for high-performance information processing have attracted
renewed interest. Neuromorphic silicon photonics has the potential to integrate
processing functions that vastly exceed the capabilities of electronics. We
report first observations of a recurrent silicon photonic neural network, in
which connections are configured by microring weight banks. A mathematical
isomorphism between the silicon photonic circuit and a continuous neural
network model is demonstrated through dynamical bifurcation analysis.
Exploiting this isomorphism, a simulated 24-node silicon photonic neural
network is programmed using "neural compiler" to solve a differential system
emulation task. A 294-fold acceleration against a conventional benchmark is
predicted. We also propose and derive power consumption analysis for
modulator-class neurons that, as opposed to laser-class neurons, are compatible
with silicon photonic platforms. At increased scale, Neuromorphic silicon
photonics could access new regimes of ultrafast information processing for
radio, control, and scientific computing.
@misc{tait2016neuromorphic,
abstract = {Photonic systems for high-performance information processing have attracted
renewed interest. Neuromorphic silicon photonics has the potential to integrate
processing functions that vastly exceed the capabilities of electronics. We
report first observations of a recurrent silicon photonic neural network, in
which connections are configured by microring weight banks. A mathematical
isomorphism between the silicon photonic circuit and a continuous neural
network model is demonstrated through dynamical bifurcation analysis.
Exploiting this isomorphism, a simulated 24-node silicon photonic neural
network is programmed using "neural compiler" to solve a differential system
emulation task. A 294-fold acceleration against a conventional benchmark is
predicted. We also propose and derive power consumption analysis for
modulator-class neurons that, as opposed to laser-class neurons, are compatible
with silicon photonic platforms. At increased scale, Neuromorphic silicon
photonics could access new regimes of ultrafast information processing for
radio, control, and scientific computing.},
added-at = {2018-01-15T21:27:55.000+0100},
author = {Tait, Alexander N. and de Lima, Thomas Ferreira and Zhou, Ellen and Wu, Allie X. and Nahmias, Mitchell A. and Shastri, Bhavin J. and Prucnal, Paul R.},
biburl = {https://www.bibsonomy.org/bibtex/2a38e1a5e7d8bee998564bcd7696fe15e/mardukasoka},
description = {Neuromorphic Silicon Photonic Networks},
doi = {10.1038/s41598-017-07754-z},
interhash = {8455e170c7319000a949717dfd54ed14},
intrahash = {a38e1a5e7d8bee998564bcd7696fe15e},
keywords = {quantumnuralnet},
note = {cite arxiv:1611.02272Comment: 12 pages, 4 figures, accepted in Scientific Reports},
timestamp = {2018-01-15T21:27:55.000+0100},
title = {Neuromorphic Silicon Photonic Networks},
url = {http://arxiv.org/abs/1611.02272},
year = 2016
}