{"id":"https://openalex.org/W7160935524","doi":"https://doi.org/10.48550/arxiv.2605.08352","title":"Convergence Analysis of Newton's Method for Neural Networks in the Overparameterized Limit","display_name":"Convergence Analysis of Newton's Method for Neural Networks in the Overparameterized Limit","publication_year":2026,"publication_date":"2026-05-08","ids":{"openalex":"https://openalex.org/W7160935524","doi":"https://doi.org/10.48550/arxiv.2605.08352"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2605.08352","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2605.08352","pdf_url":"https://arxiv.org/pdf/2605.08352","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":"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/2605.08352","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5088292594","display_name":"Konstantin Riedl","orcid":"https://orcid.org/0000-0002-2206-4334","observed_orcids":["https://orcid.org/0000-0002-2206-4334"]},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Riedl, Konstantin","raw_affiliation_strings":["Mathematical Institute University of Oxford"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Mathematical Institute University of Oxford","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103169334","display_name":"Konstantinos Spiliopoulos","orcid":"https://orcid.org/0000-0003-2021-6426","observed_orcids":["https://orcid.org/0000-0003-2021-6426"]},"institutions":[{"id":"https://openalex.org/I111088046","display_name":"Boston University","ror":"https://ror.org/05qwgg493","country_code":"US","type":"education","lineage":["https://openalex.org/I111088046"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Spiliopoulos, Konstantinos","raw_affiliation_strings":["Department of Mathematics & Statistics Boston University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Mathematics & Statistics Boston University","institution_ids":["https://openalex.org/I111088046"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5103033774","display_name":"Justin A. Sirignano","orcid":"https://orcid.org/0000-0002-0971-1349","observed_orcids":["https://orcid.org/0000-0002-0971-1349","https://orcid.org/0000-0001-6167-8710"]},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Sirignano, Justin","raw_affiliation_strings":["Mathematical Institute University of Oxford"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Mathematical Institute University of Oxford","institution_ids":["https://openalex.org/I40120149"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"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":0.9574999809265137,"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":0.9574999809265137,"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":0.034699998795986176,"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/T12814","display_name":"Gaussian Processes and Bayesian Inference","score":0.0015999999595806003,"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/newton-s-method","display_name":"Newton's method","score":0.9259999990463257},{"id":"https://openalex.org/keywords/neural-network","display_name":"artificial neural networks","score":0.7039999961853027},{"id":"https://openalex.org/keywords/infinite-width-limit","display_name":"infinite-width limit","score":0.6859999895095825},{"id":"https://openalex.org/keywords/neural-tangent-kernel","display_name":"neural tangent kernel","score":0.7009999752044678},{"id":"https://openalex.org/keywords/convergence-analysis","display_name":"convergence analysis","score":0.7129999995231628},{"id":"https://openalex.org/keywords/spectral-bias","display_name":"spectral bias","score":0.8009999990463257},{"id":"https://openalex.org/keywords/regularization-parameter-selection","display_name":"regularization parameter selection","score":0.671999990940094},{"id":"https://openalex.org/keywords/hessian-regularization","display_name":"Hessian regularization","score":0.4950000047683716},{"id":"https://openalex.org/keywords/training-dynamics","display_name":"training dynamics","score":0.6140000224113464},{"id":"https://openalex.org/keywords/exponential-convergence","display_name":"exponential convergence","score":0.49399998784065247}],"concepts":[{"id":"https://openalex.org/C203616005","wikidata":"https://www.wikidata.org/wiki/Q620495","display_name":"Hessian matrix","level":2,"score":0.8101000189781189},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.6996999979019165},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.5845999717712402},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.5067999958992004},{"id":"https://openalex.org/C158693339","wikidata":"https://www.wikidata.org/wiki/Q190524","display_name":"Eigenvalues and eigenvectors","level":2,"score":0.5016000270843506},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4934999942779541},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.49050000309944153},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.47780001163482666},{"id":"https://openalex.org/C57869625","wikidata":"https://www.wikidata.org/wiki/Q1783502","display_name":"Rate of convergence","level":3,"score":0.4724000096321106},{"id":"https://openalex.org/C34388435","wikidata":"https://www.wikidata.org/wiki/Q2267362","display_name":"Bounded function","level":2,"score":0.46059998869895935},{"id":"https://openalex.org/C151201525","wikidata":"https://www.wikidata.org/wiki/Q177239","display_name":"Limit (mathematics)","level":2,"score":0.45350000262260437},{"id":"https://openalex.org/C130367717","wikidata":"https://www.wikidata.org/wiki/Q189791","display_name":"Diagonal","level":2,"score":0.38929998874664307},{"id":"https://openalex.org/C84545080","wikidata":"https://www.wikidata.org/wiki/Q1147936","display_name":"Condition number","level":3,"score":0.3513999879360199},{"id":"https://openalex.org/C90377204","wikidata":"https://www.wikidata.org/wiki/Q1052594","display_name":"Uniform boundedness","level":3,"score":0.3379000127315521},{"id":"https://openalex.org/C138187205","wikidata":"https://www.wikidata.org/wiki/Q131251","display_name":"Tangent","level":2,"score":0.3255999982357025},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.3138999938964844},{"id":"https://openalex.org/C152442038","wikidata":"https://www.wikidata.org/wiki/Q2778212","display_name":"Tikhonov regularization","level":3,"score":0.3003999888896942},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.2928999960422516},{"id":"https://openalex.org/C145446738","wikidata":"https://www.wikidata.org/wiki/Q319913","display_name":"Convex function","level":3,"score":0.290800005197525},{"id":"https://openalex.org/C33676613","wikidata":"https://www.wikidata.org/wiki/Q13415176","display_name":"Dimension (graph theory)","level":2,"score":0.28439998626708984},{"id":"https://openalex.org/C158622935","wikidata":"https://www.wikidata.org/wiki/Q660848","display_name":"Nonlinear system","level":2,"score":0.2775999903678894},{"id":"https://openalex.org/C31441030","wikidata":"https://www.wikidata.org/wiki/Q4291882","display_name":"Least absolute deviations","level":3,"score":0.27480000257492065},{"id":"https://openalex.org/C117898588","wikidata":"https://www.wikidata.org/wiki/Q6664310","display_name":"Local convergence","level":3,"score":0.2732999920845032},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.27239999175071716},{"id":"https://openalex.org/C115680565","wikidata":"https://www.wikidata.org/wiki/Q5977448","display_name":"Gradient method","level":2,"score":0.2623000144958496},{"id":"https://openalex.org/C48216909","wikidata":"https://www.wikidata.org/wiki/Q6889156","display_name":"Modes of convergence (annotated index)","level":5,"score":0.26019999384880066},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.258899986743927},{"id":"https://openalex.org/C207467116","wikidata":"https://www.wikidata.org/wiki/Q4385666","display_name":"Inverse","level":2,"score":0.25619998574256897}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2605.08352","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2605.08352","pdf_url":"https://arxiv.org/pdf/2605.08352","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2605.08352","doi":"https://doi.org/10.48550/arxiv.2605.08352","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.08352","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2605.08352","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2605.08352","pdf_url":"https://arxiv.org/pdf/2605.08352","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G6056722365","display_name":"DMS-EPSRC: Asymptotic Analysis of Online Training Algorithms in Machine Learning: Recurrent, Graphical, and Deep Neural Networks","funder_award_id":"2311500","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320334627","display_name":"Engineering and Physical Sciences Research Council","ror":"https://ror.org/0439y7842"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7160935524.pdf","grobid_xml":"https://content.openalex.org/works/W7160935524.grobid-xml"},"referenced_works_count":8,"referenced_works":["https://openalex.org/W2947461788","https://openalex.org/W2959176742","https://openalex.org/W3133629262","https://openalex.org/W4395443493","https://openalex.org/W4399115592","https://openalex.org/W4405626889","https://openalex.org/W4416255847","https://openalex.org/W7151898343"],"related_works":["https://openalex.org/W4283069798","https://openalex.org/W2970443625","https://openalex.org/W4417299843","https://openalex.org/W3179599150","https://openalex.org/W4287182101","https://openalex.org/W2917744435","https://openalex.org/W7165636448","https://openalex.org/W3162554303","https://openalex.org/W4401023915","https://openalex.org/W4405626889"],"abstract_inverted_index":{"A":[0],"convergence":[1,54,85,114,207],"analysis":[2,214],"is":[3,86,134],"developed":[4],"for":[5,10,105,115,144,219,244,263],"the":[6,16,20,28,36,59,65,71,89,93,103,123,131,157,162,172,175,187,195,199,206,209,221,234,251,260,272,279],"regularized":[7,252],"Newton":[8,163,261],"method":[9,139,164],"training":[11,30,196,258],"neural":[12,46],"networks":[13],"(NNs)":[14],"in":[15,33,58,97,198],"overparameterized":[17,200],"limit.":[18],"As":[19],"number":[21],"of":[22,38,102,161,177,181,208,248],"hidden":[23,249],"units":[24],"tends":[25,183],"to":[26,35,70,112,140,153,184,268],"infinity,":[27],"NN":[29,66,188,235,265],"dynamics":[31,197,211],"converge":[32,141,267],"probability":[34],"solution":[37],"a":[39,44,75,166,216,230,276],"deterministic":[40],"limit":[41,201],"equation":[42],"involving":[43],"\"Newton":[45],"tangent":[47],"kernel\"":[48],"(NNTK).":[49],"Explicit":[50],"rates":[51],"characterizing":[52],"this":[53,84,178],"are":[55],"provided":[56],"and,":[57],"infinite-width":[60],"limit,":[61],"we":[62,225,241],"prove":[63,242],"that":[64,83,151,174,271],"converges":[67],"exponentially":[68],"fast":[69],"target":[72,116],"data":[73,117,145],"(i.e.,":[74],"global":[76],"minimizer":[77],"with":[78,118,146,165],"zero":[79],"loss).":[80],"We":[81],"show":[82,226],"uniform":[87],"across":[88],"frequency":[90],"spectrum,":[91],"addressing":[92],"spectral":[94],"bias":[95],"inherent":[96],"gradient":[98,106],"descent.":[99],"The":[100],"eigenvalues":[101,129],"NTK":[104],"descent":[107],"accumulate":[108],"at":[109,229],"zero,":[110,269],"leading":[111],"slow":[113],"high-frequency":[119,147],"components.":[120,148],"In":[121,239],"contrast,":[122],"NNTK":[124],"has":[125],"uniformly":[126],"lower":[127],"bounded":[128],"if":[130],"regularization":[132,222],"parameter":[133,159],"selected":[135],"appropriately,":[136],"allowing":[137],"Newton's":[138],"more":[142],"quickly":[143],"Mathematical":[149],"challenges":[150],"need":[152],"be":[154],"addressed":[155],"include":[156],"implicit":[158],"update":[160],"potentially":[167],"indefinite":[168],"Hessian":[169,253],"matrix":[170],"and":[171,259],"fact":[173],"dimension":[176],"linear":[179],"system":[180],"equations":[182],"infinity":[185],"as":[186,202,204,233,275],"width":[189,236],"grows.":[190],"This":[191],"substantially":[192],"complicates":[193],"deriving":[194],"well":[203],"proving":[205],"finite-width":[210],"thereto.":[212],"Our":[213],"identifies":[215],"scaling":[217],"formula":[218],"selecting":[220],"parameter,":[223],"which":[224],"can":[227],"vanish":[228],"suitable":[231],"rate":[232],"becomes":[237],"larger.":[238],"addition,":[240],"that,":[243],"sufficiently":[245],"large":[246],"numbers":[247],"units,":[250],"remains":[254],"positive":[255],"definite":[256],"during":[257],"updates":[262],"individual":[264],"parameters":[266],"demonstrating":[270],"model":[273],"behaves":[274],"linearization":[277],"around":[278],"initialization.":[280]},"counts_by_year":[],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2026-05-13T00:00:00"}