View in EDS PDF Full Text

Bidirectional clustering of weights for neural networks with common weights.

Saved in:
Bibliographic Details
Title: Bidirectional clustering of weights for neural networks with common weights.
Authors: Saito, Kazumi1, Nakano, Ryohei2
Source: Systems & Computers in Japan. 9/15/2007, Vol. 38 Issue 10, p46-57. 12p. 9 Charts, 2 Graphs.
Subjects: Artificial neural networks, Computer network design & construction, Computer research, Data mining, Cluster analysis (Statistics), Error functions, Polynomials
Abstract: This paper proposes a method which succinctly structures neural networks having a few thousand weights. Here structuring means weight sharing where weights in a network are divided into clusters and weights within a cluster have the same value. We newly introduce a weight sharing technique called bidirectional clustering of weights (BCW), together with second-order optimal criteria for both cluster merging and splitting. Our experiments using two artificial data sets showed that the BCW method works well to find a succinct network structure from a neural network having about 2000 weights in both regression and classification problems. © 2007 Wiley Periodicals, Inc. Syst Comp Jpn, 38(10): 46–57, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.20535 [ABSTRACT FROM AUTHOR]
Copyright of Systems & Computers in Japan is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
FullText Links:
  – Type: pdflink
Text:
  Availability: 0
Header DbId: egs
DbLabel: Engineering Source
An: 25541683
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Bidirectional clustering of weights for neural networks with common weights.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Saito%2C+Kazumi%22">Saito, Kazumi</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Nakano%2C+Ryohei%22">Nakano, Ryohei</searchLink><relatesTo>2</relatesTo>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22Systems+%26+Computers+in+Japan%22">Systems & Computers in Japan</searchLink>. 9/15/2007, Vol. 38 Issue 10, p46-57. 12p. 9 Charts, 2 Graphs.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Artificial+neural+networks%22">Artificial neural networks</searchLink><br /><searchLink fieldCode="DE" term="%22Computer+network+design+%26+construction%22">Computer network design & construction</searchLink><br /><searchLink fieldCode="DE" term="%22Computer+research%22">Computer research</searchLink><br /><searchLink fieldCode="DE" term="%22Data+mining%22">Data mining</searchLink><br /><searchLink fieldCode="DE" term="%22Cluster+analysis+%28Statistics%29%22">Cluster analysis (Statistics)</searchLink><br /><searchLink fieldCode="DE" term="%22Error+functions%22">Error functions</searchLink><br /><searchLink fieldCode="DE" term="%22Polynomials%22">Polynomials</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: This paper proposes a method which succinctly structures neural networks having a few thousand weights. Here structuring means weight sharing where weights in a network are divided into clusters and weights within a cluster have the same value. We newly introduce a weight sharing technique called bidirectional clustering of weights (BCW), together with second-order optimal criteria for both cluster merging and splitting. Our experiments using two artificial data sets showed that the BCW method works well to find a succinct network structure from a neural network having about 2000 weights in both regression and classification problems. © 2007 Wiley Periodicals, Inc. Syst Comp Jpn, 38(10): 46–57, 2007; Published online in Wiley InterScience (<URL>www.interscience.wiley.com</URL>). DOI 10.1002/scj.20535 [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of Systems & Computers in Japan is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=25541683
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1002/scj.20535
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 12
        StartPage: 46
    Subjects:
      – SubjectFull: Artificial neural networks
        Type: general
      – SubjectFull: Computer network design & construction
        Type: general
      – SubjectFull: Computer research
        Type: general
      – SubjectFull: Data mining
        Type: general
      – SubjectFull: Cluster analysis (Statistics)
        Type: general
      – SubjectFull: Error functions
        Type: general
      – SubjectFull: Polynomials
        Type: general
    Titles:
      – TitleFull: Bidirectional clustering of weights for neural networks with common weights.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Saito, Kazumi
      – PersonEntity:
          Name:
            NameFull: Nakano, Ryohei
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 15
              M: 09
              Text: 9/15/2007
              Type: published
              Y: 2007
          Identifiers:
            – Type: issn-print
              Value: 08821666
          Numbering:
            – Type: volume
              Value: 38
            – Type: issue
              Value: 10
          Titles:
            – TitleFull: Systems & Computers in Japan
              Type: main
ResultId 1