@prefix ns1: .
@prefix ns2: .
@prefix rdf: .
@prefix rdfs: .
@prefix xml: .
@prefix xsd: .
a ns1:ScholarlyArticle ;
ns1:about ,
;
ns1:author ( ) ;
ns1:citation ,
,
,
,
,
,
,
,
,
,
;
ns1:datePublished "1991-05" ;
ns1:datePublishedReg "1991-05-01" ;
ns1:description "We consider the computational complexity of learning by neural nets. We are interested in how hard it is to design appropriate neural net architectures and to train neural nets for general and specialized learning tasks. Our main result shows that the training problem for 2-cascade neural nets (which have only two non-input nodes, one of which is hidden) is N℘-complete, which implies that finding an optimal net (in terms of the number of non-input units) that is consistent with a set of examples is also N℘-complete. This result also demonstrates a surprising gap between the computational complexities of one-node (perceptron) and two-node neural net training problems, since the perceptron training problem can be solved in polynomial time by linear programming techniques. We conjecture that training a k-cascade neural net, which is a classical threshold network training problem, is also N℘-complete, for each fixed k≥2. We also show that the problem of finding an optimal perceptron (in terms of the number of non-zero weights) consistent with a set of training examples is N℘-hard. Our neural net learning model encapsulates the idea of modular neural nets, which is a popular approach to overcoming the scaling problem in training neural nets. We investigate how much easier the training problem becomes if the class of concepts to be learned is known a priori and the net architecture is allowed to be sufficiently non-optimal. Finally, we classify several neural net optimization problems within the polynomial-time hierarchy." ;
ns1:genre "research_article" ;
ns1:inLanguage "en" ;
ns1:isAccessibleForFree true ;
ns1:isPartOf [ a ns1:PublicationVolume ;
ns1:volumeNumber "6" ],
[ a ns1:PublicationIssue ;
ns1:issueNumber "3" ],
;
ns1:name "Complexity results on learning by neural nets" ;
ns1:pagination "211-230" ;
ns1:productId [ a ns1:PropertyValue ;
ns1:name "doi" ;
ns1:value "10.1007/bf00114777" ],
[ a ns1:PropertyValue ;
ns1:name "readcube_id" ;
ns1:value "b47cde7a1fbdd49a7739c168740dfec0f8fc65214f211863f97f29d6ca4129b2" ],
[ a ns1:PropertyValue ;
ns1:name "dimensions_id" ;
ns1:value "pub.1037888244" ] ;
ns1:sameAs ,
;
ns1:sdDatePublished "2019-04-15T08:50" ;
ns1:sdLicense "https://scigraph.springernature.com/explorer/license/" ;
ns1:sdPublisher [ a ns1:Organization ;
ns1:name "Springer Nature - SN SciGraph project" ] ;
ns1:url "http://link.springer.com/10.1007/BF00114777" ;
ns2:license ;
ns2:sdDataset "articles" .
a ns1:DefinedTerm ;
ns1:inDefinedTermSet ;
ns1:name "Information and Computing Sciences" .
a ns1:DefinedTerm ;
ns1:inDefinedTermSet ;
ns1:name "Computation Theory and Mathematics" .
a ns1:Periodical ;
ns1:issn "0885-6125",
"1573-0565" ;
ns1:name "Machine Learning" .
a ns1:Person ;
ns1:affiliation ;
ns1:familyName "Lin" ;
ns1:givenName "Jyh-Han" ;
ns1:sameAs .
a ns1:Person ;
ns1:affiliation ;
ns1:familyName "Vitter" ;
ns1:givenName "Jeffrey Scott" ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:CreativeWork ;
ns1:sameAs .
a ns1:Organization ;
ns1:alternateName "Brown University" ;
ns1:name "Department of Computer Science, Brown University, 02912-1910, Providence, RI" .