Why deep neural networks: A possible theoretical explanation

Chitta Baral, Olac Fuentes, Vladik Kreinovich

Research output: ResearchChapter

Abstract

In the past, the most widely used neural networks were 3-layer ones. These networks were preferred, since one of the main advantages of the biological neural networks—which motivated the use of neural networks in computing—is their parallelism, and 3-layer networks provide the largest degree of parallelism. Recently, however, it was empirically shown that, in spite of this argument, multi-layer (“deep”) neural networks leads to a much more efficient machine learning. In this paper, we provide a possible theoretical explanation for the somewhat surprising empirical success of deep networks.

LanguageEnglish (US)
Title of host publicationStudies in Systems, Decision and Control
PublisherSpringer International Publishing
Pages1-5
Number of pages5
Volume100
DOIs
StatePublished - 2018

Publication series

NameStudies in Systems, Decision and Control
Volume100
ISSN (Print)2198-4182
ISSN (Electronic)2198-4190

Fingerprint

Neural Networks
Neural networks
Deep neural networks
neural network
Parallelism
Multilayer Neural Network
Biological Networks
Machine Learning
Computing
Network layers
Learning systems
learning
Machine learning

ASJC Scopus subject areas

  • Computer Science (miscellaneous)
  • Decision Sciences (miscellaneous)
  • Economics, Econometrics and Finance (miscellaneous)
  • Automotive Engineering
  • Control and Systems Engineering
  • Control and Optimization
  • Social Sciences (miscellaneous)

Cite this

Baral, C., Fuentes, O., & Kreinovich, V. (2018). Why deep neural networks: A possible theoretical explanation. In Studies in Systems, Decision and Control (Vol. 100, pp. 1-5). (Studies in Systems, Decision and Control; Vol. 100). Springer International Publishing. DOI: 10.1007/978-3-319-61753-4_1

Why deep neural networks : A possible theoretical explanation. / Baral, Chitta; Fuentes, Olac; Kreinovich, Vladik.

Studies in Systems, Decision and Control. Vol. 100 Springer International Publishing, 2018. p. 1-5 (Studies in Systems, Decision and Control; Vol. 100).

Research output: ResearchChapter

Baral, C, Fuentes, O & Kreinovich, V 2018, Why deep neural networks: A possible theoretical explanation. in Studies in Systems, Decision and Control. vol. 100, Studies in Systems, Decision and Control, vol. 100, Springer International Publishing, pp. 1-5. DOI: 10.1007/978-3-319-61753-4_1
Baral C, Fuentes O, Kreinovich V. Why deep neural networks: A possible theoretical explanation. In Studies in Systems, Decision and Control. Vol. 100. Springer International Publishing. 2018. p. 1-5. (Studies in Systems, Decision and Control). Available from, DOI: 10.1007/978-3-319-61753-4_1
Baral, Chitta ; Fuentes, Olac ; Kreinovich, Vladik. / Why deep neural networks : A possible theoretical explanation. Studies in Systems, Decision and Control. Vol. 100 Springer International Publishing, 2018. pp. 1-5 (Studies in Systems, Decision and Control).
@inbook{8eee1a47d9be4cb69aad1d67a8ab18ad,
title = "Why deep neural networks: A possible theoretical explanation",
abstract = "In the past, the most widely used neural networks were 3-layer ones. These networks were preferred, since one of the main advantages of the biological neural networks—which motivated the use of neural networks in computing—is their parallelism, and 3-layer networks provide the largest degree of parallelism. Recently, however, it was empirically shown that, in spite of this argument, multi-layer (“deep”) neural networks leads to a much more efficient machine learning. In this paper, we provide a possible theoretical explanation for the somewhat surprising empirical success of deep networks.",
author = "Chitta Baral and Olac Fuentes and Vladik Kreinovich",
year = "2018",
doi = "10.1007/978-3-319-61753-4_1",
volume = "100",
series = "Studies in Systems, Decision and Control",
publisher = "Springer International Publishing",
pages = "1--5",
booktitle = "Studies in Systems, Decision and Control",

}

TY - CHAP

T1 - Why deep neural networks

T2 - A possible theoretical explanation

AU - Baral,Chitta

AU - Fuentes,Olac

AU - Kreinovich,Vladik

PY - 2018

Y1 - 2018

N2 - In the past, the most widely used neural networks were 3-layer ones. These networks were preferred, since one of the main advantages of the biological neural networks—which motivated the use of neural networks in computing—is their parallelism, and 3-layer networks provide the largest degree of parallelism. Recently, however, it was empirically shown that, in spite of this argument, multi-layer (“deep”) neural networks leads to a much more efficient machine learning. In this paper, we provide a possible theoretical explanation for the somewhat surprising empirical success of deep networks.

AB - In the past, the most widely used neural networks were 3-layer ones. These networks were preferred, since one of the main advantages of the biological neural networks—which motivated the use of neural networks in computing—is their parallelism, and 3-layer networks provide the largest degree of parallelism. Recently, however, it was empirically shown that, in spite of this argument, multi-layer (“deep”) neural networks leads to a much more efficient machine learning. In this paper, we provide a possible theoretical explanation for the somewhat surprising empirical success of deep networks.

UR - http://www.scopus.com/inward/record.url?scp=85029219978&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85029219978&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-61753-4_1

DO - 10.1007/978-3-319-61753-4_1

M3 - Chapter

VL - 100

T3 - Studies in Systems, Decision and Control

SP - 1

EP - 5

BT - Studies in Systems, Decision and Control

PB - Springer International Publishing

ER -