{"id":"https://openalex.org/W3090668255","doi":"https://doi.org/10.1109/ijcnn48605.2020.9207425","title":"Improving Generalization Performance of Adaptive Learning Rate by Switching from Block Diagonal Matrix Preconditioning to SGD","display_name":"Improving Generalization Performance of Adaptive Learning Rate by Switching from Block Diagonal Matrix Preconditioning to SGD","publication_year":2020,"publication_date":"2020-07-01","ids":{"openalex":"https://openalex.org/W3090668255","doi":"https://doi.org/10.1109/ijcnn48605.2020.9207425","mag":"3090668255"},"language":"en","primary_location":{"id":"doi:10.1109/ijcnn48605.2020.9207425","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ijcnn48605.2020.9207425","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 International Joint Conference on Neural Networks (IJCNN)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5086509276","display_name":"Yasutoshi Ida","orcid":"https://orcid.org/0000-0003-4279-9503"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Yasutoshi Ida","raw_affiliation_strings":["NTT Software Innovation Center, Tokyo, Japan"],"affiliations":[{"raw_affiliation_string":"NTT Software Innovation Center, Tokyo, Japan","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101917967","display_name":"Yasuhiro Fujiwara","orcid":"https://orcid.org/0000-0001-9578-1118"},"institutions":[{"id":"https://openalex.org/I2251713219","display_name":"NTT (Japan)","ror":"https://ror.org/00berct97","country_code":"JP","type":"company","lineage":["https://openalex.org/I2251713219"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Yasuhiro Fujiwara","raw_affiliation_strings":["NTT Communication Science Laboratories, Tokyo, Japan"],"affiliations":[{"raw_affiliation_string":"NTT Communication Science Laboratories, Tokyo, Japan","institution_ids":["https://openalex.org/I2251713219"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5086509276"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.2651,"has_fulltext":false,"cited_by_count":4,"citation_normalized_percentile":{"value":0.63588125,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"8"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.9997000098228455,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.9997000098228455,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9997000098228455,"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/T12676","display_name":"Machine Learning and ELM","score":0.9997000098228455,"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/generalization","display_name":"Generalization","score":0.7483172416687012},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7207714915275574},{"id":"https://openalex.org/keywords/diagonal","display_name":"Diagonal","score":0.6291373372077942},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.49999570846557617},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.46999654173851013},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.42733532190322876},{"id":"https://openalex.org/keywords/adaptive-learning","display_name":"Adaptive learning","score":0.4204123020172119},{"id":"https://openalex.org/keywords/covariance","display_name":"Covariance","score":0.41554537415504456},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.41031017899513245},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.39391061663627625},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.3725065588951111},{"id":"https://openalex.org/keywords/mathematical-optimization","display_name":"Mathematical optimization","score":0.3340323567390442},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.18447351455688477}],"concepts":[{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.7483172416687012},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7207714915275574},{"id":"https://openalex.org/C130367717","wikidata":"https://www.wikidata.org/wiki/Q189791","display_name":"Diagonal","level":2,"score":0.6291373372077942},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.49999570846557617},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.46999654173851013},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.42733532190322876},{"id":"https://openalex.org/C125014702","wikidata":"https://www.wikidata.org/wiki/Q4680749","display_name":"Adaptive learning","level":2,"score":0.4204123020172119},{"id":"https://openalex.org/C178650346","wikidata":"https://www.wikidata.org/wiki/Q201984","display_name":"Covariance","level":2,"score":0.41554537415504456},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.41031017899513245},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.39391061663627625},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3725065588951111},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3340323567390442},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.18447351455688477},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/ijcnn48605.2020.9207425","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ijcnn48605.2020.9207425","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 International Joint Conference on Neural Networks (IJCNN)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":72,"referenced_works":["https://openalex.org/W6908809","https://openalex.org/W155854392","https://openalex.org/W1522301498","https://openalex.org/W1686810756","https://openalex.org/W1772464306","https://openalex.org/W1970789124","https://openalex.org/W1994616650","https://openalex.org/W2033245860","https://openalex.org/W2116456623","https://openalex.org/W2142426721","https://openalex.org/W2146502635","https://openalex.org/W2148991744","https://openalex.org/W2155894447","https://openalex.org/W2160208155","https://openalex.org/W2163605009","https://openalex.org/W2170240176","https://openalex.org/W2407521645","https://openalex.org/W2525778437","https://openalex.org/W2561368111","https://openalex.org/W2619368999","https://openalex.org/W2619516334","https://openalex.org/W2740560623","https://openalex.org/W2750384547","https://openalex.org/W2753431380","https://openalex.org/W2769856846","https://openalex.org/W2776855315","https://openalex.org/W2785523195","https://openalex.org/W2785940809","https://openalex.org/W2804386825","https://openalex.org/W2950800384","https://openalex.org/W2951841718","https://openalex.org/W2962820688","https://openalex.org/W2962915600","https://openalex.org/W2963012544","https://openalex.org/W2963208657","https://openalex.org/W2963241285","https://openalex.org/W2963586744","https://openalex.org/W2963712897","https://openalex.org/W2964121744","https://openalex.org/W2964125128","https://openalex.org/W2971154439","https://openalex.org/W2977665223","https://openalex.org/W3102242707","https://openalex.org/W3118608800","https://openalex.org/W4229633200","https://openalex.org/W4295181861","https://openalex.org/W6600284362","https://openalex.org/W6606341887","https://openalex.org/W6631190155","https://openalex.org/W6637373629","https://openalex.org/W6637940777","https://openalex.org/W6677681515","https://openalex.org/W6681435938","https://openalex.org/W6681804681","https://openalex.org/W6683107984","https://openalex.org/W6684191040","https://openalex.org/W6685053522","https://openalex.org/W6693969136","https://openalex.org/W6714138976","https://openalex.org/W6730419752","https://openalex.org/W6736221594","https://openalex.org/W6738491991","https://openalex.org/W6743688258","https://openalex.org/W6744061804","https://openalex.org/W6746514494","https://openalex.org/W6747043858","https://openalex.org/W6747620207","https://openalex.org/W6748482152","https://openalex.org/W6751983209","https://openalex.org/W6765987481","https://openalex.org/W6767676143","https://openalex.org/W6787972765"],"related_works":["https://openalex.org/W2135584473","https://openalex.org/W3162204513","https://openalex.org/W2964223869","https://openalex.org/W2792539140","https://openalex.org/W4375867731","https://openalex.org/W2371138613","https://openalex.org/W2005808740","https://openalex.org/W2295130398","https://openalex.org/W2048963458","https://openalex.org/W43109613"],"abstract_inverted_index":{"Deep":[0],"Neural":[1],"Networks":[2],"(DNNs)":[3],"are":[4,15,41],"widely":[5],"used":[6,93],"for":[7,17],"various":[8],"applications.":[9],"Although":[10],"adaptive":[11,55,96,110],"learning":[12,56,97,111],"rate":[13,57,98,112],"algorithms":[14,99,113],"attractive":[16],"DNN":[18,44],"training,":[19],"their":[20,102],"theoretical":[21],"performance":[22,118],"remains":[23],"unclear.":[24],"In":[25,104],"fact,":[26],"published":[27],"analyses":[28],"consider":[29],"only":[30],"simple":[31],"optimization":[32],"settings":[33],"such":[34],"as":[35],"convex":[36],"optimization,":[37],"none":[38],"of":[39,66,84,123],"which":[40],"applicable":[42],"to":[43],"training.":[45,124],"This":[46],"paper":[47],"proposes":[48],"TSO-ALRA,":[49],"a":[50,63,136],"two-stage":[51],"optimizer":[52],"using":[53],"an":[54,131],"algorithm;":[58],"it":[59],"is":[60],"based":[61],"on":[62,77,141],"full":[64],"analysis":[65,87,107],"two":[67],"approaches":[68],"that":[69,89,109,147],"do":[70],"suit":[71],"DNNs:":[72],"parameter":[73],"updates":[74],"along":[75],"geodesics":[76],"the":[78,90,120],"statistical":[79],"manifold":[80],"and":[81,135,144],"covariance":[82],"structure":[83],"gradients.":[85],"Our":[86,139],"reveals":[88],"diagonal":[91],"approximation":[92,133],"by":[94],"existing":[95],"inevitably":[100],"degrades":[101],"efficiency.":[103],"addition,":[105],"our":[106],"suggests":[108],"suffer":[114],"drops":[115],"in":[116,119],"generalization":[117,153],"last":[121],"phase":[122],"To":[125],"overcome":[126],"these":[127],"problems,":[128],"TSO-ALRA":[129,148],"combines":[130],"effective":[132],"technique":[134],"switching":[137],"strategy.":[138],"experiments":[140],"several":[142],"models":[143],"datasets":[145],"show":[146],"efficiently":[149],"converges":[150],"with":[151],"high":[152],"performance.":[154]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2022,"cited_by_count":1},{"year":2021,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}