{"id":"https://openalex.org/W7212030746","doi":"https://doi.org/10.48550/arxiv.2609.09130","title":"Nearly Tight Rademacher Bounds for Sparsely Activated Neural Networks","display_name":"Nearly Tight Rademacher Bounds for Sparsely Activated Neural Networks","publication_year":2026,"publication_date":"2026-09-08","ids":{"openalex":"https://openalex.org/W7212030746","doi":"https://doi.org/10.48550/arxiv.2609.09130"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.09130","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.09130","pdf_url":"https://arxiv.org/pdf/2609.09130","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.09130","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5144125913","display_name":"Xiaoyu Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I31746571","display_name":"UNSW Sydney","ror":"https://ror.org/03r8z3t63","country_code":"AU","type":"education","lineage":["https://openalex.org/I31746571"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Li, Xiaoyu","raw_affiliation_strings":["University of New South Wales"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of New South Wales","institution_ids":["https://openalex.org/I31746571"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110738653","display_name":"Zhizhou Sha","orcid":"https://orcid.org/0009-0004-2494-1945","observed_orcids":["https://orcid.org/0009-0004-2494-1945"]},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sha, Zhizhou","raw_affiliation_strings":["University of Texas at Austin"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Texas at Austin","institution_ids":["https://openalex.org/I86519309"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5151071890","display_name":"Jiaojiao Jiang","orcid":"https://orcid.org/0000-0001-7307-8114","observed_orcids":["https://orcid.org/0000-0001-7307-8114"]},"institutions":[{"id":"https://openalex.org/I31746571","display_name":"UNSW Sydney","ror":"https://ror.org/03r8z3t63","country_code":"AU","type":"education","lineage":["https://openalex.org/I31746571"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Jiang, Jiaojiao","raw_affiliation_strings":["University of New South Wales"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of New South Wales","institution_ids":["https://openalex.org/I31746571"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133309459","display_name":"Junbin Gao","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I129604602","display_name":"The University of Sydney","ror":"https://ror.org/0384j8v12","country_code":"AU","type":"education","lineage":["https://openalex.org/I129604602"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Gao, Junbin","raw_affiliation_strings":["University of Sydney"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Sydney","institution_ids":["https://openalex.org/I129604602"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5031625303","display_name":"Andi Han","orcid":"https://orcid.org/0000-0003-4655-655X","observed_orcids":["https://orcid.org/0000-0003-4655-655X"]},"institutions":[{"id":"https://openalex.org/I129604602","display_name":"The University of Sydney","ror":"https://ror.org/0384j8v12","country_code":"AU","type":"education","lineage":["https://openalex.org/I129604602"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Han, Andi","raw_affiliation_strings":["University of Sydney"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Sydney","institution_ids":["https://openalex.org/I129604602"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.7470999956130981,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.7470999956130981,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.1843000054359436,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12072","display_name":"Machine Learning and Algorithms","score":0.04769999906420708,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/rademacher-complexity","display_name":"Rademacher complexity","score":0.9169999957084656},{"id":"https://openalex.org/keywords/relu-neural-networks","display_name":"ReLU networks","score":0.722000002861023},{"id":"https://openalex.org/keywords/statistical-complexity","display_name":"statistical complexity","score":0.5199999809265137},{"id":"https://openalex.org/keywords/lower-bounds","display_name":"lower bounds","score":0.6330000162124634},{"id":"https://openalex.org/keywords/one-hidden-layer-neural-networks","display_name":"one-hidden-layer neural networks","score":0.621999979019165},{"id":"https://openalex.org/keywords/width-dependence","display_name":"width dependence","score":0.39899998903274536}],"concepts":[{"id":"https://openalex.org/C39927690","wikidata":"https://www.wikidata.org/wiki/Q11197","display_name":"Logarithm","level":2,"score":0.6310999989509583},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.6007000207901001},{"id":"https://openalex.org/C34388435","wikidata":"https://www.wikidata.org/wiki/Q2267362","display_name":"Bounded function","level":2,"score":0.5985999703407288},{"id":"https://openalex.org/C49020025","wikidata":"https://www.wikidata.org/wiki/Q1059099","display_name":"Chaining","level":2,"score":0.5580999851226807},{"id":"https://openalex.org/C119322782","wikidata":"https://www.wikidata.org/wiki/Q2662236","display_name":"VC dimension","level":2,"score":0.5439000129699707},{"id":"https://openalex.org/C2780428219","wikidata":"https://www.wikidata.org/wiki/Q16952335","display_name":"Cover (algebra)","level":2,"score":0.4936999976634979},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.4733000099658966},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4569999873638153},{"id":"https://openalex.org/C77553402","wikidata":"https://www.wikidata.org/wiki/Q13222579","display_name":"Upper and lower bounds","level":2,"score":0.4544999897480011},{"id":"https://openalex.org/C33676613","wikidata":"https://www.wikidata.org/wiki/Q13415176","display_name":"Dimension (graph theory)","level":2,"score":0.45100000500679016},{"id":"https://openalex.org/C118615104","wikidata":"https://www.wikidata.org/wiki/Q121416","display_name":"Discrete mathematics","level":1,"score":0.436599999666214},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4196999967098236},{"id":"https://openalex.org/C149728462","wikidata":"https://www.wikidata.org/wiki/Q751319","display_name":"Minimax","level":2,"score":0.39980000257492065},{"id":"https://openalex.org/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.3919999897480011},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.3723999857902527},{"id":"https://openalex.org/C97256817","wikidata":"https://www.wikidata.org/wiki/Q1462316","display_name":"Spurious relationship","level":2,"score":0.3481999933719635},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.33880001306533813},{"id":"https://openalex.org/C122041747","wikidata":"https://www.wikidata.org/wiki/Q838611","display_name":"Ball (mathematics)","level":2,"score":0.33880001306533813},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.33570000529289246},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.3248000144958496},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.30329999327659607},{"id":"https://openalex.org/C2778445095","wikidata":"https://www.wikidata.org/wiki/Q18354077","display_name":"Sample complexity","level":2,"score":0.30070000886917114},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.28600001335144043},{"id":"https://openalex.org/C129848803","wikidata":"https://www.wikidata.org/wiki/Q2564360","display_name":"Sample size determination","level":2,"score":0.28279998898506165},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.26809999346733093},{"id":"https://openalex.org/C183115368","wikidata":"https://www.wikidata.org/wiki/Q856577","display_name":"Weighting","level":2,"score":0.2644999921321869},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.2572000026702881},{"id":"https://openalex.org/C46681722","wikidata":"https://www.wikidata.org/wiki/Q4725589","display_name":"Alias","level":2,"score":0.2526000142097473}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.09130","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.09130","pdf_url":"https://arxiv.org/pdf/2609.09130","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.09130","doi":"https://doi.org/10.48550/arxiv.2609.09130","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.09130","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.09130","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.09130","pdf_url":"https://arxiv.org/pdf/2609.09130","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7212030746.pdf","grobid_xml":"https://content.openalex.org/works/W7212030746.grobid-xml"},"referenced_works_count":2,"referenced_works":["https://openalex.org/W1511694993","https://openalex.org/W4251232133"],"related_works":["https://openalex.org/W2777138330","https://openalex.org/W4391598006","https://openalex.org/W4313483655","https://openalex.org/W4318751297","https://openalex.org/W7021247103","https://openalex.org/W3038677001","https://openalex.org/W7114807493","https://openalex.org/W4383173863","https://openalex.org/W3139459228","https://openalex.org/W4309399130"],"abstract_inverted_index":{"An":[0],"input":[1,61,108],"may":[2],"activate":[3],"few":[4],"hidden":[5],"units":[6,43,100,123],"even":[7],"when":[8],"different":[9],"inputs":[10,102],"collectively":[11],"use":[12],"an":[13],"entire":[14,116],"network.":[15],"We":[16],"study":[17],"the":[18,26,57,77,115,135,150,158],"statistical":[19],"complexity":[20,125],"of":[21,30,179],"this":[22],"input-dependent":[23],"sparsity":[24,155],"in":[25,56,142],"one-hidden-layer":[27],"ReLU":[28],"model":[29],"Awasthi":[31],"et":[32],"al.":[33],"(COLT":[34],"2024).":[35],"For":[36,161],"width":[37,105],"$s$,":[38],"at":[39,119],"most":[40,120],"$k$":[41],"active":[42,99],"per":[44],"input,":[45],"and":[46,49,70,124,167],"effective":[47],"weight":[48],"bias":[50,129],"bounds":[51,86,130,178],"$W,B$,":[52],"every":[53],"size-$m$":[54],"sample":[55],"class's":[58],"fixed":[59],"radius-$R$":[60],"domain":[62,109,141],"satisfies":[63],"$\\mathcal{R}(S)\\le":[64],"CWR\\min\\{k,\\sqrt{sk/m}\\log^{3/2}(2m)\\}+kB/\\sqrt":[65],"m$.":[66],"A":[67,146],"support-preserving":[68],"cover":[69],"a":[71,104,162],"single":[72],"normalized":[73,164],"chaining":[74],"argument":[75],"remove":[76],"previous":[78],"explicit":[79],"dimension":[80],"factor,":[81],"up":[82,92,182],"to":[83,93,132,170,183],"logarithms.":[84,184],"Lower":[85],"on":[87,114,138,157],"appropriate":[88],"i.i.d.":[89],"marginals":[90],"match":[91],"those":[94],"logarithms,":[95],"showing":[96],"how":[97],"changing":[98],"across":[101],"retains":[103],"dependence.":[106],"The":[107],"matters:":[110],"zero-bias":[111],"networks":[112],"sparse":[113],"ball":[117],"have":[118],"$2k$":[121],"nonzero":[122],"$O(kWR/\\sqrt":[126],"m)$,":[127],"whereas":[128],"comparable":[131,169],"$WR$":[133],"restore":[134],"worst-case":[136],"rate":[137],"that":[139],"same":[140],"only":[143],"logarithmic":[144],"dimension.":[145],"spherical-cap":[147],"construction":[148],"proves":[149],"latter":[151],"claim":[152],"without":[153],"assuming":[154],"merely":[156],"sampling":[159],"support.":[160],"specified":[163],"bounded":[165],"loss":[166],"biases":[168],"$WR$,":[171],"we":[172],"also":[173],"obtain":[174],"agnostic":[175],"minimax":[176],"excess-risk":[177],"order":[180],"$\\min\\{1,\\sqrt{s/(km)}\\}$":[181]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-10T00:00:00"}