{"id":"https://openalex.org/W3205186138","doi":null,"title":"ES-Based Jacobian Enables Faster Bilevel Optimization","display_name":"ES-Based Jacobian Enables Faster Bilevel Optimization","publication_year":2021,"publication_date":"2021-10-13","ids":{"openalex":"https://openalex.org/W3205186138","mag":"3205186138"},"language":"en","primary_location":{"id":"mag:3205186138","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/pdf/2110.07004.pdf","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"http://arxiv.org/pdf/2110.07004.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5030734394","display_name":"Daouda Sow","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Daouda Sow","raw_affiliation_strings":["Ohio State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5071973105","display_name":"Kaiyi Ji","orcid":"https://orcid.org/0000-0002-9533-0058","observed_orcids":["https://orcid.org/0000-0002-9533-0058"]},"institutions":[{"id":"https://openalex.org/I27837315","display_name":"University of Michigan","ror":"https://ror.org/00jmfr291","country_code":"US","type":"education","lineage":["https://openalex.org/I27837315"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kaiyi Ji","raw_affiliation_strings":["\u2020University of Michigan"],"raw_orcid":"https://orcid.org/0000-0002-9533-0058","affiliations":[{"raw_affiliation_string":"\u2020University of Michigan","institution_ids":["https://openalex.org/I27837315"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100384384","display_name":"Yingbin Liang","orcid":"https://orcid.org/0000-0003-2631-4262","observed_orcids":["https://orcid.org/0000-0003-2631-4262","https://orcid.org/0000-0002-8635-2992"]},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yingbin Liang","raw_affiliation_strings":["Ohio State University"],"raw_orcid":"https://orcid.org/0000-0003-2631-4262","affiliations":[{"raw_affiliation_string":"Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]}],"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":false,"cited_by_count":1,"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":0.9682999849319458,"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":0.9682999849319458,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.01679999940097332,"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":0.0020000000949949026,"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/bi-level-optimization","display_name":"bi-level optimization","score":1.0},{"id":"https://openalex.org/keywords/evolution-strategies","display_name":"evolution strategies","score":0.9739999771118164},{"id":"https://openalex.org/keywords/jacobian-approximation","display_name":"Jacobian approximation","score":0.7940000295639038},{"id":"https://openalex.org/keywords/stochastic-bilevel-optimization","display_name":"stochastic bilevel optimization","score":0.7260000109672546},{"id":"https://openalex.org/keywords/convergence-analysis","display_name":"convergence analysis","score":0.6980000138282776},{"id":"https://openalex.org/keywords/computational-complexity","display_name":"computational complexity","score":0.8799999952316284}],"concepts":[{"id":"https://openalex.org/C200331156","wikidata":"https://www.wikidata.org/wiki/Q506041","display_name":"Jacobian matrix and determinant","level":2,"score":0.8928015232086182},{"id":"https://openalex.org/C3309286","wikidata":"https://www.wikidata.org/wiki/Q4907693","display_name":"Bilevel optimization","level":3,"score":0.6713541746139526},{"id":"https://openalex.org/C203616005","wikidata":"https://www.wikidata.org/wiki/Q620495","display_name":"Hessian matrix","level":2,"score":0.6426403522491455},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6177824139595032},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5527914762496948},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.5394836664199829},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5032927393913269},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.48393183946609497},{"id":"https://openalex.org/C133512626","wikidata":"https://www.wikidata.org/wiki/Q787371","display_name":"Automatic differentiation","level":3,"score":0.4552142322063446},{"id":"https://openalex.org/C89109886","wikidata":"https://www.wikidata.org/wiki/Q1535924","display_name":"Trust region","level":3,"score":0.4352002739906311},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4251977503299713},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.399901807308197},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.3603551983833313},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.33153069019317627},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2891981601715088},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.18901896476745605},{"id":"https://openalex.org/C178635117","wikidata":"https://www.wikidata.org/wiki/Q747499","display_name":"RADIUS","level":2,"score":0.0},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0},{"id":"https://openalex.org/C50522688","wikidata":"https://www.wikidata.org/wiki/Q189833","display_name":"Economic growth","level":1,"score":0.0},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.0},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"mag:3205186138","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/pdf/2110.07004.pdf","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null}],"best_oa_location":{"id":"mag:3205186138","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/pdf/2110.07004.pdf","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":49,"referenced_works":["https://openalex.org/W102487131","https://openalex.org/W1983202079","https://openalex.org/W2031264011","https://openalex.org/W2112796928","https://openalex.org/W2117539524","https://openalex.org/W2149479912","https://openalex.org/W2187061624","https://openalex.org/W2270245739","https://openalex.org/W2287409372","https://openalex.org/W2505728881","https://openalex.org/W2601450892","https://openalex.org/W2604763608","https://openalex.org/W2787415863","https://openalex.org/W2803895400","https://openalex.org/W2810075754","https://openalex.org/W2888451463","https://openalex.org/W2907670343","https://openalex.org/W2914953695","https://openalex.org/W2934413201","https://openalex.org/W2945204933","https://openalex.org/W2963233958","https://openalex.org/W2963243330","https://openalex.org/W2963306862","https://openalex.org/W2963341924","https://openalex.org/W2963470657","https://openalex.org/W2963804140","https://openalex.org/W2963852457","https://openalex.org/W2964121744","https://openalex.org/W2970697704","https://openalex.org/W2974885182","https://openalex.org/W2995627493","https://openalex.org/W3008360155","https://openalex.org/W3034220177","https://openalex.org/W3035016095","https://openalex.org/W3037120332","https://openalex.org/W3046916097","https://openalex.org/W3100702524","https://openalex.org/W3123350987","https://openalex.org/W3126652544","https://openalex.org/W3127438322","https://openalex.org/W3136431828","https://openalex.org/W3158974429","https://openalex.org/W3167489592","https://openalex.org/W3167655307","https://openalex.org/W3169203126","https://openalex.org/W3170327376","https://openalex.org/W3170720259","https://openalex.org/W3212367831","https://openalex.org/W4317897806"],"related_works":["https://openalex.org/W7133206543","https://openalex.org/W3112047150","https://openalex.org/W3092809425","https://openalex.org/W7133206818","https://openalex.org/W3170720259","https://openalex.org/W3167655307","https://openalex.org/W3193127553","https://openalex.org/W7061487028","https://openalex.org/W4377865264","https://openalex.org/W4400064923"],"abstract_inverted_index":{"Bilevel":[0],"optimization":[1],"(BO)":[2],"has":[3],"arisen":[4],"as":[5,87,104],"a":[6,54,155],"powerful":[7],"tool":[8],"for":[9,91,115],"solving":[10],"many":[11],"modern":[12],"machine":[13],"learning":[14],"problems.":[15,138],"However,":[16],"due":[17],"to":[18,65,100,128],"the":[19,67,72,92,101,109,121,129,132,144,150,159,165,176],"nested":[20],"structure":[21],"of":[22,74,123,131,154,167],"BO,":[23,75],"existing":[24],"gradient-based":[25],"methods":[26,134],"require":[27],"second-order":[28,81],"derivative":[29],"approximations":[30],"via":[31],"Jacobian-":[32],"or/and":[33],"Hessian-vector":[34],"computations,":[35],"which":[36,58,162],"can":[37],"be":[38],"very":[39],"costly":[40],"in":[41,71,140,143,175],"practice,":[42],"especially":[43],"with":[44],"large":[45],"neural":[46],"network":[47,157],"models.":[48],"In":[49],"this":[50],"work,":[51],"we":[52,107,119,148],"propose":[53],"novel":[55],"BO":[56],"algorithm,":[57],"adopts":[59],"Evolution":[60],"Strategies":[61],"(ES)":[62],"based":[63],"method":[64],"approximate":[66],"response":[68],"Jacobian":[69,94],"matrix":[70],"hypergradient":[73],"and":[76,96,112,172],"hence":[77],"fully":[78],"eliminates":[79],"all":[80],"computations.":[82],"We":[83],"call":[84],"our":[85,116,124,141,168],"algorithm":[86],"ESJ":[88],"(which":[89],"stands":[90],"ES-based":[93,169],"method)":[95],"further":[97],"extend":[98],"it":[99],"stochastic":[102],"setting":[103],"ESJ-S.":[105],"Theoretically,":[106],"characterize":[108],"convergence":[110],"guarantee":[111],"computational":[113],"complexity":[114],"algorithms.":[117],"Experimentally,":[118],"demonstrate":[120],"superiority":[122],"proposed":[125],"algorithms":[126],"compared":[127],"state":[130],"art":[133],"on":[135],"various":[136],"bilevel":[137,170],"Particularly,":[139],"experiment":[142],"few-shot":[145],"meta-learning":[146],"problem,":[147],"meta-learn":[149],"twelve":[151],"millions":[152],"parameters":[153],"ResNet-12":[156],"over":[158],"miniImageNet":[160],"dataset,":[161],"evidently":[163],"demonstrates":[164],"scalability":[166],"approach":[171],"its":[173],"feasibility":[174],"large-scale":[177],"setting.":[178]},"counts_by_year":[{"year":2023,"cited_by_count":1}],"updated_date":"2026-11-04T10:54:05.112534","created_date":"2021-10-25T00:00:00"}