{"id":"https://openalex.org/W7214035937","doi":"https://doi.org/10.48550/arxiv.2609.24522","title":"Convergence of Rotation-based Matrix Optimizers: A Unified Analysis of SOAP, Conda, and SPlus","display_name":"Convergence of Rotation-based Matrix Optimizers: A Unified Analysis of SOAP, Conda, and SPlus","publication_year":2026,"publication_date":"2026-09-21","ids":{"openalex":"https://openalex.org/W7214035937","doi":"https://doi.org/10.48550/arxiv.2609.24522"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.24522","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.24522","pdf_url":"https://arxiv.org/pdf/2609.24522","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/2609.24522","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5152478645","display_name":"Yiwen Sun","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sun, Yiwen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5122971972","display_name":"Huan Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I205237279","display_name":"Nankai University","ror":"https://ror.org/01y1kjr75","country_code":"CN","type":"education","lineage":["https://openalex.org/I205237279"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Li, Huan","raw_affiliation_strings":["Institute of Robotics and Automatic Information Systems , College of Artificial Intelligence , Nankai University , Tianjin , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Institute of Robotics and Automatic Information Systems , College of Artificial Intelligence , Nankai University , Tianjin , China","institution_ids":["https://openalex.org/I205237279"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5122963047","display_name":"Zhouchen Lin","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lin, Zhouchen","raw_affiliation_strings":["National Key Lab of General AI , School of Intelligence Science and Technology , Peking University , Beijing , China. "],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Key Lab of General AI , School of Intelligence Science and Technology , Peking University , Beijing , China. ","institution_ids":["https://openalex.org/I20231570"]}]}],"institutions":[],"countries_distinct_count":1,"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":0.9976000189781189,"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":0.9976000189781189,"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/T10792","display_name":"Matrix Theory and Algorithms","score":0.0010999999940395355,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/T10500","display_name":"Sparse and Compressive Sensing Techniques","score":0.0003000000142492354,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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/matrix-optimization","display_name":"matrix optimization","score":0.7120000123977661},{"id":"https://openalex.org/keywords/convergence-analysis","display_name":"convergence analysis","score":0.7310000061988831},{"id":"https://openalex.org/keywords/adam-optimizer","display_name":"Adam optimizer","score":0.7300000190734863},{"id":"https://openalex.org/keywords/orthogonal-transformations","display_name":"orthogonal transformations","score":0.8130000233650208},{"id":"https://openalex.org/keywords/dimension-dependence","display_name":"dimension dependence","score":0.5199999809265137},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"stochastic gradient descent","score":0.42399999499320984}],"concepts":[{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.6940000057220459},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.6324999928474426},{"id":"https://openalex.org/C33676613","wikidata":"https://www.wikidata.org/wiki/Q13415176","display_name":"Dimension (graph theory)","level":2,"score":0.6308000087738037},{"id":"https://openalex.org/C57869625","wikidata":"https://www.wikidata.org/wiki/Q1783502","display_name":"Rate of convergence","level":3,"score":0.6248999834060669},{"id":"https://openalex.org/C130367717","wikidata":"https://www.wikidata.org/wiki/Q189791","display_name":"Diagonal","level":2,"score":0.5881999731063843},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.5662000179290771},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.5109999775886536},{"id":"https://openalex.org/C148043351","wikidata":"https://www.wikidata.org/wiki/Q4456944","display_name":"Current (fluid)","level":2,"score":0.4081999957561493},{"id":"https://openalex.org/C113313756","wikidata":"https://www.wikidata.org/wiki/Q332791","display_name":"Diagonal matrix","level":3,"score":0.3921999931335449},{"id":"https://openalex.org/C74050887","wikidata":"https://www.wikidata.org/wiki/Q848368","display_name":"Rotation (mathematics)","level":2,"score":0.39079999923706055},{"id":"https://openalex.org/C83633838","wikidata":"https://www.wikidata.org/wiki/Q1256564","display_name":"Rotation matrix","level":2,"score":0.38350000977516174},{"id":"https://openalex.org/C48216909","wikidata":"https://www.wikidata.org/wiki/Q6889156","display_name":"Modes of convergence (annotated index)","level":5,"score":0.3727000057697296},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3707999885082245},{"id":"https://openalex.org/C54848796","wikidata":"https://www.wikidata.org/wiki/Q339011","display_name":"Symmetric matrix","level":3,"score":0.3659000098705292},{"id":"https://openalex.org/C44292817","wikidata":"https://www.wikidata.org/wiki/Q333871","display_name":"Orthogonal matrix","level":3,"score":0.3467999994754791},{"id":"https://openalex.org/C49555168","wikidata":"https://www.wikidata.org/wiki/Q176583","display_name":"Stochastic matrix","level":3,"score":0.3434000015258789},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.33889999985694885},{"id":"https://openalex.org/C165443888","wikidata":"https://www.wikidata.org/wiki/Q1482183","display_name":"Transformation matrix","level":3,"score":0.32710000872612},{"id":"https://openalex.org/C5871375","wikidata":"https://www.wikidata.org/wiki/Q1780715","display_name":"Compact convergence","level":4,"score":0.30640000104904175},{"id":"https://openalex.org/C2780813799","wikidata":"https://www.wikidata.org/wiki/Q3274237","display_name":"Zero (linguistics)","level":2,"score":0.3025999963283539},{"id":"https://openalex.org/C64812099","wikidata":"https://www.wikidata.org/wiki/Q176604","display_name":"Random matrix","level":3,"score":0.29789999127388},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.27889999747276306},{"id":"https://openalex.org/C2777299769","wikidata":"https://www.wikidata.org/wiki/Q3707858","display_name":"Type (biology)","level":2,"score":0.2694000005722046},{"id":"https://openalex.org/C149488123","wikidata":"https://www.wikidata.org/wiki/Q17098730","display_name":"Matrix analysis","level":3,"score":0.26809999346733093},{"id":"https://openalex.org/C14928880","wikidata":"https://www.wikidata.org/wiki/Q1512559","display_name":"Normal convergence","level":4,"score":0.26750001311302185},{"id":"https://openalex.org/C2780428219","wikidata":"https://www.wikidata.org/wiki/Q16952335","display_name":"Cover (algebra)","level":2,"score":0.25699999928474426},{"id":"https://openalex.org/C98184364","wikidata":"https://www.wikidata.org/wiki/Q1780131","display_name":"Argument (complex analysis)","level":2,"score":0.25279998779296875}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.24522","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.24522","pdf_url":"https://arxiv.org/pdf/2609.24522","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.2609.24522","doi":"https://doi.org/10.48550/arxiv.2609.24522","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.24522","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:2609.24522","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.24522","pdf_url":"https://arxiv.org/pdf/2609.24522","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":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214035937.pdf","grobid_xml":"https://content.openalex.org/works/W7214035937.grobid-xml"},"referenced_works_count":5,"referenced_works":["https://openalex.org/W4407212433","https://openalex.org/W4415331769","https://openalex.org/W4417156118","https://openalex.org/W7083821503","https://openalex.org/W7128499965"],"related_works":["https://openalex.org/W7155531991","https://openalex.org/W7213300824","https://openalex.org/W7125923009","https://openalex.org/W7215000185","https://openalex.org/W7214751144","https://openalex.org/W7164927518","https://openalex.org/W4407183549","https://openalex.org/W7129320626","https://openalex.org/W7166714282","https://openalex.org/W4416895734"],"abstract_inverted_index":{"In":[0],"this":[1],"work,":[2],"we":[3,66],"develop":[4],"a":[5,26,59,98,128],"unified":[6],"theoretical":[7],"framework":[8,42,95,126],"for":[9,68],"analyzing":[10],"the":[11,23,38,69,72,84,91,117,145,149],"convergence":[12,73,85],"of":[13,75,87,102,148],"rotation-based":[14,103],"matrix":[15,45],"optimizers,":[16,46],"which":[17],"apply":[18],"orthogonal":[19,108],"transformations":[20],"to":[21,114],"map":[22],"momentum":[24],"into":[25,136],"rotated":[27],"space,":[28],"perform":[29],"coordinate-wise":[30],"or":[31],"normalized":[32],"updates":[33,39],"there,":[34],"and":[35,50],"then":[36],"rotate":[37],"back.":[40],"Our":[41,94],"encompasses":[43],"prominent":[44],"including":[47],"SOAP,":[48],"Conda,":[49],"truncated":[51],"SPlus,":[52],"as":[53,55,58,81,83],"well":[54,82],"matrix-parameterized":[56],"Adam":[57,88],"special":[60],"case.":[61],"As":[62],"our":[63,125],"main":[64],"result,":[65],"establish,":[67],"first":[70],"time,":[71],"rate":[74,86],"SOAP":[76],"with":[77],"sharp":[78],"dimensional":[79],"dependence,":[80],"measured":[89],"by":[90],"nuclear":[92],"norm.":[93],"also":[96],"covers":[97],"more":[99],"general":[100],"variant":[101],"optimizers":[104],"that":[105,132],"allows":[106],"arbitrary":[107],"rotation":[109],"matrices,":[110,142],"allowing":[111],"these":[112],"matrices":[113],"depend":[115],"on":[116,138],"current":[118],"stochastic":[119],"gradient":[120],"at":[121],"each":[122],"iteration.":[123],"Technically,":[124],"leverages":[127],"row-column":[129],"trace-control":[130],"argument":[131],"converts":[133],"elementwise":[134],"bounds":[135,137],"two":[139],"diagonal":[140],"control":[141],"thereby":[143],"improving":[144],"dimension":[146],"dependence":[147],"resulting":[150],"bound.":[151]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-23T00:00:00"}