Hardware efficient learning on a 3-D optoelectronic neural system

Ashok V. Krishnamoorthy; Stephen A. Brodsky; Clark C. Guest; Gary C. Marsden; Matthias Blume; Gokce Yayla; Jean Merckle; Sadik C. Esener

Hardware efficient learning on a 3-D optoelectronic neural system

Ashok V. Krishnamoorthy, Stephen A. Brodsky, Clark C. Guest, Gary C. Marsden, Matthias Blume, Gokce Yayla, Jean Merckle, Sadik C. Esener

Research output: Contribution to conference › Paper › peer-review

1 Scopus citations

Abstract

We discuss the Dual-Scale Topology Optoelectronic Processor (D-STOP) neural network, a scalable, optically interconnected neural network architecture. We present the tandem D-STOP system, which provides the connectivity needed for building fully-parallel neural networks with generic gradient-descent learning rules. We review the Content Addressable Network (CAN) learning algorithm, a discrete learning algorithm that provides accelerated learning with reduced hardware requirements. We then show how the CAN algorithm can be effectively mapped onto D-STOP, and we investigate associated optoelectronic hardware tradeoffs.

Original language	English (US)
Pages	1998-2003
Number of pages	6
State	Published - 1994
Externally published	Yes
Event	Proceedings of the 1994 IEEE International Conference on Neural Networks. Part 1 (of 7) - Orlando, FL, USA Duration: Jun 27 1994 → Jun 29 1994

Other

Other	Proceedings of the 1994 IEEE International Conference on Neural Networks. Part 1 (of 7)
City	Orlando, FL, USA
Period	6/27/94 → 6/29/94

ASJC Scopus subject areas

Software

Cite this

@conference{8e927d2f462b4534926a56921db68d4e,

title = "Hardware efficient learning on a 3-D optoelectronic neural system",

abstract = "We discuss the Dual-Scale Topology Optoelectronic Processor (D-STOP) neural network, a scalable, optically interconnected neural network architecture. We present the tandem D-STOP system, which provides the connectivity needed for building fully-parallel neural networks with generic gradient-descent learning rules. We review the Content Addressable Network (CAN) learning algorithm, a discrete learning algorithm that provides accelerated learning with reduced hardware requirements. We then show how the CAN algorithm can be effectively mapped onto D-STOP, and we investigate associated optoelectronic hardware tradeoffs.",

author = "Krishnamoorthy, {Ashok V.} and Brodsky, {Stephen A.} and Guest, {Clark C.} and Marsden, {Gary C.} and Matthias Blume and Gokce Yayla and Jean Merckle and Esener, {Sadik C.}",

year = "1994",

language = "English (US)",

pages = "1998--2003",

note = "Proceedings of the 1994 IEEE International Conference on Neural Networks. Part 1 (of 7) ; Conference date: 27-06-1994 Through 29-06-1994",

}

TY - CONF

T1 - Hardware efficient learning on a 3-D optoelectronic neural system

AU - Krishnamoorthy, Ashok V.

AU - Brodsky, Stephen A.

AU - Guest, Clark C.

AU - Marsden, Gary C.

AU - Blume, Matthias

AU - Yayla, Gokce

AU - Merckle, Jean

AU - Esener, Sadik C.

PY - 1994

Y1 - 1994

N2 - We discuss the Dual-Scale Topology Optoelectronic Processor (D-STOP) neural network, a scalable, optically interconnected neural network architecture. We present the tandem D-STOP system, which provides the connectivity needed for building fully-parallel neural networks with generic gradient-descent learning rules. We review the Content Addressable Network (CAN) learning algorithm, a discrete learning algorithm that provides accelerated learning with reduced hardware requirements. We then show how the CAN algorithm can be effectively mapped onto D-STOP, and we investigate associated optoelectronic hardware tradeoffs.

AB - We discuss the Dual-Scale Topology Optoelectronic Processor (D-STOP) neural network, a scalable, optically interconnected neural network architecture. We present the tandem D-STOP system, which provides the connectivity needed for building fully-parallel neural networks with generic gradient-descent learning rules. We review the Content Addressable Network (CAN) learning algorithm, a discrete learning algorithm that provides accelerated learning with reduced hardware requirements. We then show how the CAN algorithm can be effectively mapped onto D-STOP, and we investigate associated optoelectronic hardware tradeoffs.

UR - http://www.scopus.com/inward/record.url?scp=0028752593&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0028752593&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:0028752593

SP - 1998

EP - 2003

T2 - Proceedings of the 1994 IEEE International Conference on Neural Networks. Part 1 (of 7)

Y2 - 27 June 1994 through 29 June 1994

ER -

Hardware efficient learning on a 3-D optoelectronic neural system

Abstract

Other

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this