{"id":"https://openalex.org/W7167916285","doi":"https://doi.org/10.48550/arxiv.2607.08104","title":"Vanilla SGD with Momentum Survives Heavy-Tailed Noise: Convergence Analysis without Gradient Clipping or Normalization","display_name":"Vanilla SGD with Momentum Survives Heavy-Tailed Noise: Convergence Analysis without Gradient Clipping or Normalization","publication_year":2026,"publication_date":"2026-07-09","ids":{"openalex":"https://openalex.org/W7167916285","doi":"https://doi.org/10.48550/arxiv.2607.08104"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.08104","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.08104","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,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2607.08104","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5140381936","display_name":"Ryusei Yamada","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yamada, Ryusei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5107467232","display_name":"N Sato","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sato, Naoki","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5016609320","display_name":"Hideaki Iiduka","orcid":"https://orcid.org/0000-0001-9173-6723"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Iiduka, Hideaki","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"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":0.9861000180244446,"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":0.9861000180244446,"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/T11413","display_name":"Risk and Portfolio Optimization","score":0.001500000013038516,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T11195","display_name":"Simulation Techniques and Applications","score":0.00139999995008111,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/normalization","display_name":"Normalization (sociology)","score":0.6805999875068665},{"id":"https://openalex.org/keywords/clipping","display_name":"Clipping (morphology)","score":0.6363000273704529},{"id":"https://openalex.org/keywords/convergence","display_name":"Convergence (economics)","score":0.5582000017166138},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"Stochastic gradient descent","score":0.5465999841690063},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.5365999937057495},{"id":"https://openalex.org/keywords/momentum","display_name":"Momentum (technical analysis)","score":0.5062999725341797}],"concepts":[{"id":"https://openalex.org/C136886441","wikidata":"https://www.wikidata.org/wiki/Q926129","display_name":"Normalization (sociology)","level":2,"score":0.6805999875068665},{"id":"https://openalex.org/C2776848632","wikidata":"https://www.wikidata.org/wiki/Q853463","display_name":"Clipping (morphology)","level":2,"score":0.6363000273704529},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.5921000242233276},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.5582000017166138},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.5465999841690063},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.5365999937057495},{"id":"https://openalex.org/C60718061","wikidata":"https://www.wikidata.org/wiki/Q1414747","display_name":"Momentum (technical analysis)","level":2,"score":0.5062999725341797},{"id":"https://openalex.org/C57869625","wikidata":"https://www.wikidata.org/wiki/Q1783502","display_name":"Rate of convergence","level":3,"score":0.4526999890804291},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.44850000739097595},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.365200012922287},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3424000144004822},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3303999900817871},{"id":"https://openalex.org/C115680565","wikidata":"https://www.wikidata.org/wiki/Q5977448","display_name":"Gradient method","level":2,"score":0.301800012588501},{"id":"https://openalex.org/C203616005","wikidata":"https://www.wikidata.org/wiki/Q620495","display_name":"Hessian matrix","level":2,"score":0.3010999858379364},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.26989999413490295}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.08104","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.08104","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,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2607.08104","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.08104","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,"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"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Stochastic":[0],"gradient":[1,24,83],"descent":[2],"(SGD)":[3],"is":[4,16],"a":[5,30],"cornerstone":[6],"of":[7,45,68,106,112],"modern":[8],"optimization.":[9],"While":[10],"its":[11],"performance":[12],"under":[13,115],"heavy-tailed":[14,46,116],"noise":[15],"often":[17],"addressed":[18],"through":[19],"specialized":[20],"modifications":[21],"such":[22],"as":[23],"clipping":[25],"or":[26,103],"normalization,":[27],"we":[28,51],"investigate":[29],"more":[31,60],"fundamental":[32],"question:":[33],"how":[34],"does":[35],"vanilla":[36,57,69,113],"SGD,":[37,107],"particularly":[38],"with":[39,71],"momentum,":[40],"perform":[41],"in":[42],"the":[43,63,90,97],"presence":[44],"noise?":[47],"In":[48],"this":[49],"paper,":[50],"refine":[52],"existing":[53],"convergence":[54,66,92],"results":[55,87],"for":[56,73],"SGD":[58,70],"and,":[59],"importantly,":[61],"provide":[62],"first":[64],"comprehensive":[65],"analysis":[67],"momentum":[72],"strongly":[74],"convex,":[75,76],"and":[77],"nonconvex":[78],"objectives,":[79],"without":[80],"employing":[81],"any":[82],"control":[84],"mechanisms.":[85],"Our":[86],"demonstrate":[88],"that":[89],"obtained":[91],"rates":[93,99],"are":[94,121],"inferior":[95],"to":[96],"optimal":[98],"achieved":[100],"by":[101,123],"clipped":[102],"normalized":[104],"variants":[105],"thereby":[108],"revealing":[109],"inherent":[110],"limitations":[111],"methods":[114],"noise.":[117],"The":[118],"theoretical":[119],"findings":[120],"supported":[122],"experiments":[124],"on":[125],"synthetic":[126],"functions.":[127]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-07-11T00:00:00"}