Benefit of deep learning with non-convex noisy gradient descent: Provable excess risk bound and superiority to kernel methods. | D2R Server publishing the DBLP Bibliography Database, hosted at L3S Research Center

Property	Value
dcterms:bibliographicCitation	<http://dblp.uni-trier.de/rec/bibtex/conf/iclr/SuzukiA21>
dc:creator	<https://dblp.l3s.de/d2r/resource/authors/Shunta_Akiyama>
dc:creator	<https://dblp.l3s.de/d2r/resource/authors/Taiji_Suzuki>
foaf:homepage	<https://openreview.net/forum?id=2m0g1wEafh>
dc:identifier	DBLP conf/iclr/SuzukiA21 (xsd:string)
dcterms:issued	2021 (xsd:gYear)
rdfs:label	Benefit of deep learning with non-convex noisy gradient descent: Provable excess risk bound and superiority to kernel methods. (xsd:string)
foaf:maker	<https://dblp.l3s.de/d2r/resource/authors/Shunta_Akiyama>
foaf:maker	<https://dblp.l3s.de/d2r/resource/authors/Taiji_Suzuki>
dcterms:partOf	<https://dblp.l3s.de/d2r/resource/publications/conf/iclr/2021>
owl:sameAs	<http://bibsonomy.org/uri/bibtexkey/conf/iclr/SuzukiA21/dblp>
owl:sameAs	<http://dblp.rkbexplorer.com/id/conf/iclr/SuzukiA21>
rdfs:seeAlso	<http://dblp.uni-trier.de/db/conf/iclr/iclr2021.html#SuzukiA21>
rdfs:seeAlso	<https://openreview.net/forum?id=2m0g1wEafh>
swrc:series	<https://dblp.l3s.de/d2r/resource/conferences/iclr>
dc:title	Benefit of deep learning with non-convex noisy gradient descent: Provable excess risk bound and superiority to kernel methods. (xsd:string)
dc:type	<http://purl.org/dc/dcmitype/Text>
rdf:type	swrc:InProceedings
rdf:type	foaf:Document