{"id":"https://openalex.org/W4387929635","doi":"https://doi.org/10.48550/arxiv.2310.14963","title":"Studying K-FAC Heuristics by Viewing Adam through a Second-Order Lens","display_name":"Studying K-FAC Heuristics by Viewing Adam through a Second-Order Lens","publication_year":2023,"publication_date":"2023-10-23","ids":{"openalex":"https://openalex.org/W4387929635","doi":"https://doi.org/10.48550/arxiv.2310.14963"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2310.14963","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2310.14963","pdf_url":"https://arxiv.org/pdf/2310.14963","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/2310.14963","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5039277324","display_name":"Ross M. Clarke","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Clarke, Ross M.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5102890597","display_name":"Jos\u00e9 Miguel Hern\u00e1ndez-Lobato","orcid":"https://orcid.org/0000-0001-7610-949X","observed_orcids":["https://orcid.org/0000-0001-7610-949X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hern\u00e1ndez-Lobato, Jos\u00e9 Miguel","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":true,"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":0.9984999895095825,"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":0.9984999895095825,"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":0.00019999999494757503,"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":0.00019999999494757503,"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/k-fac","display_name":"K-FAC","score":0.9700000286102295},{"id":"https://openalex.org/keywords/learning-rate-selection","display_name":"learning rate selection","score":0.6660000085830688},{"id":"https://openalex.org/keywords/hyperparameter-optimization","display_name":"hyperparameter optimization","score":0.7739999890327454},{"id":"https://openalex.org/keywords/quasi-newton-methods","display_name":"quasi-Newton methods","score":0.9210000038146973}],"concepts":[{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.6753849983215332},{"id":"https://openalex.org/C182306322","wikidata":"https://www.wikidata.org/wiki/Q1779371","display_name":"Order (exchange)","level":2,"score":0.5951886177062988},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5604457855224609},{"id":"https://openalex.org/C195065555","wikidata":"https://www.wikidata.org/wiki/Q214881","display_name":"Curvature","level":2,"score":0.5442959070205688},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.5355980396270752},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5296573042869568},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.47909262776374817},{"id":"https://openalex.org/C43091099","wikidata":"https://www.wikidata.org/wiki/Q1067788","display_name":"Through-the-lens metering","level":3,"score":0.4709119498729706},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4334523677825928},{"id":"https://openalex.org/C15336307","wikidata":"https://www.wikidata.org/wiki/Q1766051","display_name":"Lens (geology)","level":2,"score":0.40746742486953735},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3618200421333313},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3209115266799927},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2613843083381653},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.12023600935935974},{"id":"https://openalex.org/C120665830","wikidata":"https://www.wikidata.org/wiki/Q14620","display_name":"Optics","level":1,"score":0.10185077786445618},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.08774816989898682},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.08430653810501099},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.08348238468170166},{"id":"https://openalex.org/C10138342","wikidata":"https://www.wikidata.org/wiki/Q43015","display_name":"Finance","level":1,"score":0.0},{"id":"https://openalex.org/C146978453","wikidata":"https://www.wikidata.org/wiki/Q3798668","display_name":"Aerospace engineering","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2310.14963","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2310.14963","pdf_url":"https://arxiv.org/pdf/2310.14963","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.2310.14963","doi":"https://doi.org/10.48550/arxiv.2310.14963","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2310.14963","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:2310.14963","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2310.14963","pdf_url":"https://arxiv.org/pdf/2310.14963","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/G3887492883","display_name":null,"funder_award_id":"2107369","funder_id":"https://openalex.org/F4320334627","funder_display_name":"Engineering and Physical Sciences Research Council"},{"id":"https://openalex.org/G6738147560","display_name":"Turing AI Fellowship: Machine Learning for Molecular Design","funder_award_id":"EP/V023756/1","funder_id":"https://openalex.org/F4320334627","funder_display_name":"Engineering and Physical Sciences Research Council"},{"id":"https://openalex.org/G8402139738","display_name":null,"funder_award_id":"EP/S515334/1","funder_id":"https://openalex.org/F4320334627","funder_display_name":"Engineering and Physical Sciences Research Council"}],"funders":[{"id":"https://openalex.org/F4320317153","display_name":"DeepMind","ror":"https://ror.org/00971b260"},{"id":"https://openalex.org/F4320334627","display_name":"Engineering and Physical Sciences Research Council","ror":"https://ror.org/0439y7842"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4387929635.pdf","grobid_xml":"https://content.openalex.org/works/W4387929635.grobid-xml"},"referenced_works_count":2,"referenced_works":["https://openalex.org/W1518461880","https://openalex.org/W4365211469"],"related_works":["https://openalex.org/W3004050881","https://openalex.org/W3031420959","https://openalex.org/W2996323231","https://openalex.org/W2980368140","https://openalex.org/W4405873075","https://openalex.org/W4417167354","https://openalex.org/W4386609751","https://openalex.org/W4226333858","https://openalex.org/W4200634098","https://openalex.org/W4391556643"],"abstract_inverted_index":{"Research":[0],"into":[1],"optimisation":[2],"for":[3],"deep":[4],"learning":[5,82],"is":[6],"characterised":[7],"by":[8,97,100],"a":[9,104,111],"tension":[10],"between":[11],"the":[12,25,47,64,93],"computational":[13],"efficiency":[14,27],"of":[15,28,49,113,130],"first-order,":[16],"gradient-based":[17],"methods":[18,31,35,41],"(such":[19,32,52],"as":[20,33,53],"SGD":[21],"and":[22,24,36,81,115,121,135],"Adam)":[23],"theoretical":[26],"second-order,":[29],"curvature-based":[30],"quasi-Newton":[34],"K-FAC).":[37],"Noting":[38],"that":[39],"second-order":[40,65,70,105],"often":[42],"only":[43],"function":[44],"effectively":[45],"with":[46,92],"addition":[48],"stabilising":[50],"heuristics":[51,128],"Levenberg-Marquardt":[54],"damping),":[55],"we":[56],"ask":[57],"how":[58],"much":[59],"these":[60],"(as":[61],"opposed":[62],"to":[63,69,147],"curvature":[66],"model)":[67],"contribute":[68],"algorithms'":[71],"performance.":[72],"We":[73,107],"thus":[74],"study":[75],"AdamQLR:":[76],"an":[77,137],"optimiser":[78],"combining":[79],"damping":[80],"rate":[83],"selection":[84],"techniques":[85],"from":[86],"K-FAC":[87],"(Martens":[88],"&":[89],"Grosse,":[90],"2015)":[91],"update":[94],"directions":[95],"proposed":[96],"Adam,":[98],"inspired":[99],"considering":[101],"Adam":[102],"through":[103],"lens.":[106],"evaluate":[108],"AdamQLR":[109,139],"on":[110],"range":[112],"regression":[114],"classification":[116],"tasks":[117],"at":[118],"various":[119],"scales":[120],"hyperparameter":[122],"tuning":[123],"methodologies,":[124],"concluding":[125],"K-FAC's":[126],"adaptive":[127],"are":[129],"variable":[131],"standalone":[132],"general":[133],"effectiveness,":[134],"finding":[136],"untuned":[138],"setting":[140],"can":[141],"achieve":[142],"comparable":[143],"performance":[144],"vs":[145],"runtime":[146],"tuned":[148],"benchmarks.":[149]},"counts_by_year":[{"year":2024,"cited_by_count":1}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2023-10-25T00:00:00"}