{"id":"https://openalex.org/W4403815243","doi":"https://doi.org/10.48550/arxiv.2409.20325","title":"Old Optimizer, New Norm: An Anthology","display_name":"Old Optimizer, New Norm: An Anthology","publication_year":2024,"publication_date":"2024-09-30","ids":{"openalex":"https://openalex.org/W4403815243","doi":"https://doi.org/10.48550/arxiv.2409.20325"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2409.20325","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2409.20325","pdf_url":"https://arxiv.org/pdf/2409.20325","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/2409.20325","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5067399219","display_name":"Jeremy Bernstein","orcid":"https://orcid.org/0000-0001-9110-7476","observed_orcids":["https://orcid.org/0000-0001-9110-7476"]},"institutions":[{"id":"https://openalex.org/I4405258931","display_name":"MIT Computer Science and Artificial Intelligence Laboratory","ror":"https://ror.org/01e85r662","country_code":"US","type":"facility","lineage":["https://openalex.org/I4405258931","https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bernstein, Jeremy","raw_affiliation_strings":["MIT CSAIL , United States"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"MIT CSAIL , United States","institution_ids":["https://openalex.org/I4405258931"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5114425769","display_name":"Laker Newhouse","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405258931","display_name":"MIT Computer Science and Artificial Intelligence Laboratory","ror":"https://ror.org/01e85r662","country_code":"US","type":"facility","lineage":["https://openalex.org/I4405258931","https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Newhouse, Laker","raw_affiliation_strings":["MIT CSAIL , United States"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"MIT CSAIL , United States","institution_ids":["https://openalex.org/I4405258931"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I4405258931"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":200,"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":0.9990000128746033,"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":0.9990000128746033,"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":0.00039999998989515007,"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/T12535","display_name":"Machine Learning and Data Classification","score":9.999999747378752e-05,"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/adam-optimizer","display_name":"Adam optimizer","score":0.8019999861717224},{"id":"https://openalex.org/keywords/first-order-methods","display_name":"first-order methods","score":0.6620000004768372},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"gradient descent","score":0.8309999704360962},{"id":"https://openalex.org/keywords/operator-norm","display_name":"operator norm","score":0.8980000019073486},{"id":"https://openalex.org/keywords/exponentially-weighted-moving-average","display_name":"exponential moving average","score":0.7350000143051147},{"id":"https://openalex.org/keywords/training-stability","display_name":"training stability","score":0.39100000262260437}],"concepts":[{"id":"https://openalex.org/C191795146","wikidata":"https://www.wikidata.org/wiki/Q3878446","display_name":"Norm (philosophy)","level":2,"score":0.6523500084877014},{"id":"https://openalex.org/C144237770","wikidata":"https://www.wikidata.org/wiki/Q747534","display_name":"Mathematical economics","level":1,"score":0.5153169631958008},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.38516145944595337},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.363986611366272},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.2253512144088745},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.20720577239990234}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2409.20325","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2409.20325","pdf_url":"https://arxiv.org/pdf/2409.20325","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.2409.20325","doi":"https://doi.org/10.48550/arxiv.2409.20325","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2409.20325","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:2409.20325","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2409.20325","pdf_url":"https://arxiv.org/pdf/2409.20325","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/W4403815243.pdf","grobid_xml":"https://content.openalex.org/works/W4403815243.grobid-xml"},"referenced_works_count":11,"referenced_works":["https://openalex.org/W2746647004","https://openalex.org/W3133629262","https://openalex.org/W4365211469","https://openalex.org/W4380375175","https://openalex.org/W4380559103","https://openalex.org/W4385507657","https://openalex.org/W4386793277","https://openalex.org/W4388090547","https://openalex.org/W4400065274","https://openalex.org/W4400611281","https://openalex.org/W4415796879"],"related_works":["https://openalex.org/W4417100728","https://openalex.org/W7134925487","https://openalex.org/W7163210198","https://openalex.org/W4415796879","https://openalex.org/W2737818580","https://openalex.org/W3129487256","https://openalex.org/W3035418911","https://openalex.org/W3113161081","https://openalex.org/W7092314007","https://openalex.org/W3087978365"],"abstract_inverted_index":{"Deep":[0],"learning":[1],"optimizers":[2],"are":[3],"often":[4],"motivated":[5],"through":[6],"a":[7,36,59,68],"mix":[8],"of":[9,108,127],"convex":[10],"and":[11,23,26,99,116,139],"approximate":[12],"second-order":[13],"theory.":[14],"We":[15,122],"select":[16],"three":[17],"such":[18],"methods":[19],"--":[20,25],"Adam,":[21],"Shampoo":[22],"Prodigy":[24],"argue":[27],"that":[28,88,124],"each":[29,51],"method":[30,39,52],"can":[31],"instead":[32],"be":[33,79,118],"understood":[34],"as":[35],"squarely":[37],"first-order":[38],"without":[40],"convexity":[41],"assumptions.":[42],"In":[43],"fact,":[44],"after":[45],"switching":[46],"off":[47],"exponential":[48],"moving":[49],"averages,":[50],"is":[53],"equivalent":[54],"to":[55,81,135],"steepest":[56],"descent":[57],"under":[58],"particular":[60],"norm.":[61],"By":[62],"generalizing":[63],"this":[64,125],"observation,":[65],"we":[66],"chart":[67],"new":[69],"design":[70],"space":[71,107],"for":[72],"training":[73],"algorithms.":[74],"Different":[75],"operator":[76],"norms":[77],"should":[78,117],"assigned":[80,119],"different":[82,114,120],"tensors":[83],"based":[84],"on":[85],"the":[86,89,93,104,130],"role":[87],"tensor":[90],"plays":[91],"within":[92],"network.":[94],"For":[95],"example,":[96],"while":[97],"linear":[98],"embedding":[100],"layers":[101,112],"may":[102],"have":[103],"same":[105],"weight":[106],"$\\mathbb{R}^{m\\times":[109],"n}$,":[110],"these":[111],"play":[113],"roles":[115],"norms.":[121],"hope":[123],"idea":[126],"carefully":[128],"metrizing":[129],"neural":[131],"architecture":[132],"might":[133],"lead":[134],"more":[136],"stable,":[137],"scalable":[138],"indeed":[140],"faster":[141],"training.":[142]},"counts_by_year":[{"year":2026,"cited_by_count":142},{"year":2025,"cited_by_count":54},{"year":2024,"cited_by_count":3},{"year":2023,"cited_by_count":1}],"updated_date":"2026-11-12T12:28:57.059296","created_date":"2024-10-28T00:00:00"}