{"id":"https://openalex.org/W4404942737","doi":"https://doi.org/10.1088/2632-2153/ad9a39","title":"Stabilizing training of affine coupling layers for high-dimensional variational inference","display_name":"Stabilizing training of affine coupling layers for high-dimensional variational inference","publication_year":2024,"publication_date":"2024-12-01","ids":{"openalex":"https://openalex.org/W4404942737","doi":"https://doi.org/10.1088/2632-2153/ad9a39"},"language":"en","primary_location":{"id":"doi:10.1088/2632-2153/ad9a39","is_oa":true,"landing_page_url":"https://doi.org/10.1088/2632-2153/ad9a39","pdf_url":"https://iopscience.iop.org/article/10.1088/2632-2153/ad9a39/pdf","source":{"id":"https://openalex.org/S4210200687","display_name":"Machine Learning Science and Technology","issn_l":"2632-2153","issn":["2632-2153"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320083","host_organization_name":"IOP Publishing","host_organization_lineage":["https://openalex.org/P4310320083","https://openalex.org/P4310311669"],"host_organization_lineage_names":["IOP Publishing","Institute of Physics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Machine Learning: Science and Technology","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","doaj"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://iopscience.iop.org/article/10.1088/2632-2153/ad9a39/pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101506514","display_name":"Daniel Andrade","orcid":"https://orcid.org/0000-0002-1123-4369"},"institutions":[{"id":"https://openalex.org/I57930482","display_name":"Hiroshima City University","ror":"https://ror.org/001et4e78","country_code":"JP","type":"education","lineage":["https://openalex.org/I57930482"]}],"countries":["JP"],"is_corresponding":true,"raw_author_name":"Daniel Andrade","raw_affiliation_strings":["School of Informatics and Data Science, Hiroshima University, Higashi-Hiroshima City, Hiroshima, Japan"],"raw_orcid":"https://orcid.org/0000-0002-1123-4369","affiliations":[{"raw_affiliation_string":"School of Informatics and Data Science, Hiroshima University, Higashi-Hiroshima City, Hiroshima, Japan","institution_ids":["https://openalex.org/I57930482"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":["https://openalex.org/A5101506514"],"corresponding_institution_ids":["https://openalex.org/I57930482"],"apc_list":{"value":1600,"currency":"GBP","value_usd":1962},"apc_paid":{"value":1600,"currency":"GBP","value_usd":1962},"fwci":0.2044,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.51189831,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":95},"biblio":{"volume":"5","issue":"4","first_page":"045066","last_page":"045066"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.9990000128746033,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.9990000128746033,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T12814","display_name":"Gaussian Processes and Bayesian Inference","score":0.9962000250816345,"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/T11206","display_name":"Model Reduction and Neural Networks","score":0.9940999746322632,"subfield":{"id":"https://openalex.org/subfields/3109","display_name":"Statistical and Nonlinear Physics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/affine-transformation","display_name":"Affine transformation","score":0.7235552072525024},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.7065185904502869},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.6063552498817444},{"id":"https://openalex.org/keywords/coupling","display_name":"Coupling (piping)","score":0.5667387247085571},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.49054259061813354},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4175998866558075},{"id":"https://openalex.org/keywords/applied-mathematics","display_name":"Applied mathematics","score":0.3419877886772156},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.3339947760105133},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.18792837858200073},{"id":"https://openalex.org/keywords/pure-mathematics","display_name":"Pure mathematics","score":0.18294087052345276},{"id":"https://openalex.org/keywords/physics","display_name":"Physics","score":0.17870381474494934},{"id":"https://openalex.org/keywords/mechanical-engineering","display_name":"Mechanical engineering","score":0.13089457154273987}],"concepts":[{"id":"https://openalex.org/C92757383","wikidata":"https://www.wikidata.org/wiki/Q382497","display_name":"Affine transformation","level":2,"score":0.7235552072525024},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7065185904502869},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.6063552498817444},{"id":"https://openalex.org/C131584629","wikidata":"https://www.wikidata.org/wiki/Q4308705","display_name":"Coupling (piping)","level":2,"score":0.5667387247085571},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.49054259061813354},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4175998866558075},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.3419877886772156},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3339947760105133},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.18792837858200073},{"id":"https://openalex.org/C202444582","wikidata":"https://www.wikidata.org/wiki/Q837863","display_name":"Pure mathematics","level":1,"score":0.18294087052345276},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.17870381474494934},{"id":"https://openalex.org/C78519656","wikidata":"https://www.wikidata.org/wiki/Q101333","display_name":"Mechanical engineering","level":1,"score":0.13089457154273987},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1088/2632-2153/ad9a39","is_oa":true,"landing_page_url":"https://doi.org/10.1088/2632-2153/ad9a39","pdf_url":"https://iopscience.iop.org/article/10.1088/2632-2153/ad9a39/pdf","source":{"id":"https://openalex.org/S4210200687","display_name":"Machine Learning Science and Technology","issn_l":"2632-2153","issn":["2632-2153"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320083","host_organization_name":"IOP Publishing","host_organization_lineage":["https://openalex.org/P4310320083","https://openalex.org/P4310311669"],"host_organization_lineage_names":["IOP Publishing","Institute of Physics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Machine Learning: Science and Technology","raw_type":"journal-article"},{"id":"pmh:oai:doaj.org/article:c1b8fbdcd66c433387d43adb4fd17a89","is_oa":true,"landing_page_url":"https://doaj.org/article/c1b8fbdcd66c433387d43adb4fd17a89","pdf_url":null,"source":{"id":"https://openalex.org/S4306401280","display_name":"DOAJ (DOAJ: Directory of Open Access Journals)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Machine Learning: Science and Technology, Vol 5, Iss 4, p 045066 (2024)","raw_type":"article"}],"best_oa_location":{"id":"doi:10.1088/2632-2153/ad9a39","is_oa":true,"landing_page_url":"https://doi.org/10.1088/2632-2153/ad9a39","pdf_url":"https://iopscience.iop.org/article/10.1088/2632-2153/ad9a39/pdf","source":{"id":"https://openalex.org/S4210200687","display_name":"Machine Learning Science and Technology","issn_l":"2632-2153","issn":["2632-2153"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320083","host_organization_name":"IOP Publishing","host_organization_lineage":["https://openalex.org/P4310320083","https://openalex.org/P4310311669"],"host_organization_lineage_names":["IOP Publishing","Institute of Physics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Machine Learning: Science and Technology","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3021056888","display_name":"Computationally Feasible Confidence Regions for Bayes Factors with Iteratively Refined Normalizing Flows","funder_award_id":"22K11934","funder_id":"https://openalex.org/F4320334764","funder_display_name":"Japan Society for the Promotion of Science"}],"funders":[{"id":"https://openalex.org/F4320334764","display_name":"Japan Society for the Promotion of Science","ror":"https://ror.org/00hhkn466"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4404942737.pdf","grobid_xml":"https://content.openalex.org/works/W4404942737.grobid-xml"},"referenced_works_count":77,"referenced_works":["https://openalex.org/W32980360","https://openalex.org/W1513873506","https://openalex.org/W1522301498","https://openalex.org/W1666386643","https://openalex.org/W1959608418","https://openalex.org/W1982652137","https://openalex.org/W2033582774","https://openalex.org/W2087684630","https://openalex.org/W2114169935","https://openalex.org/W2225156818","https://openalex.org/W2898631838","https://openalex.org/W2950985821","https://openalex.org/W2953194650","https://openalex.org/W2954340342","https://openalex.org/W2962851448","https://openalex.org/W2962994101","https://openalex.org/W2963090522","https://openalex.org/W2963173382","https://openalex.org/W2963755523","https://openalex.org/W2981150028","https://openalex.org/W3034727712","https://openalex.org/W3044820246","https://openalex.org/W3074977478","https://openalex.org/W3082269577","https://openalex.org/W3082274400","https://openalex.org/W3091324939","https://openalex.org/W3129907242","https://openalex.org/W3150807214","https://openalex.org/W3166268148","https://openalex.org/W3186199082","https://openalex.org/W3212191040","https://openalex.org/W3212685810","https://openalex.org/W4211177544","https://openalex.org/W4224910142","https://openalex.org/W4280520103","https://openalex.org/W4283316023","https://openalex.org/W4283514718","https://openalex.org/W4287331428","https://openalex.org/W4287756229","https://openalex.org/W4287867252","https://openalex.org/W4296907727","https://openalex.org/W4297798428","https://openalex.org/W4318171169","https://openalex.org/W4381888413","https://openalex.org/W4389488975","https://openalex.org/W4393241997","https://openalex.org/W4393851281","https://openalex.org/W6610566761","https://openalex.org/W6631190155","https://openalex.org/W6637133718","https://openalex.org/W6640963894","https://openalex.org/W6682648773","https://openalex.org/W6696727771","https://openalex.org/W6714644935","https://openalex.org/W6742563595","https://openalex.org/W6752307458","https://openalex.org/W6755609140","https://openalex.org/W6759750808","https://openalex.org/W6763486065","https://openalex.org/W6764817228","https://openalex.org/W6769403462","https://openalex.org/W6770847983","https://openalex.org/W6773951134","https://openalex.org/W6779301693","https://openalex.org/W6779585612","https://openalex.org/W6782469754","https://openalex.org/W6784345087","https://openalex.org/W6790424659","https://openalex.org/W6790965610","https://openalex.org/W6791352959","https://openalex.org/W6796743115","https://openalex.org/W6804057142","https://openalex.org/W6810789937","https://openalex.org/W6838136636","https://openalex.org/W6838990407","https://openalex.org/W6839557214","https://openalex.org/W6862835396"],"related_works":["https://openalex.org/W230091440","https://openalex.org/W2233261550","https://openalex.org/W2810751659","https://openalex.org/W258997015","https://openalex.org/W2997094352","https://openalex.org/W3216976533","https://openalex.org/W100620283","https://openalex.org/W2495260952","https://openalex.org/W4366179611","https://openalex.org/W2996078371"],"abstract_inverted_index":{"Abstract":[0],"Variational":[1],"inference":[2],"with":[3,168,178],"normalizing":[4,16,40,54,235],"flows":[5,17,41,55,236],"is":[6,61,185],"an":[7],"increasingly":[8],"popular":[9],"alternative":[10],"to":[11,29,44,65,92,238],"MCMC":[12],"methods.":[13],"In":[14,34,73],"particular,":[15],"based":[18],"on":[19,152],"affine":[20],"coupling":[21],"layers":[22],"(Real":[23],"NVPs)":[24],"are":[25],"frequently":[26],"used":[27],"due":[28,64],"their":[30],"good":[31,239],"empirical":[32,224],"performance.":[33],"theory,":[35],"increasing":[36],"the":[37,66,70,83,100,103,129,142,230],"depth":[38],"of":[39,69,85,96,102,123,128,141,173,234],"should":[42],"lead":[43],"more":[45,189],"accurate":[46,190],"posterior":[47,59,240],"approximations.":[48],"However,":[49],"in":[50,131],"practice,":[51],"training":[52,95,172,202,212],"deep":[53],"for":[56,81,176,188,211,215],"approximating":[57],"high-dimensional":[58,159,216],"distributions":[60],"often":[62,111],"infeasible":[63],"high":[67,114],"variance":[68,84],"stochastic":[71,86],"gradients.":[72],"this":[74],"work,":[75],"we":[76,105,119,198,220],"show":[77,166],"that":[78,167,228],"previous":[79],"methods":[80],"stabilizing":[82],"gradient":[87],"descent":[88],"can":[89],"be":[90],"insufficient":[91],"achieve":[93],"stable":[94,171],"Real":[97,132,174,213],"NVPs.":[98],"As":[99,116],"source":[101],"problem,":[104],"identify":[106],"that,":[107],"during":[108],"training,":[109],"samples":[110],"exhibit":[112],"unusual":[113],"values.":[115],"a":[117,121,136,158],"remedy,":[118],"propose":[120],"combination":[122],"two":[124],"methods:":[125],"(1)":[126],"soft-thresholding":[127],"scale":[130],"NVPs,":[133],"and":[134,147,182,204,207,225],"(2)":[135],"bijective":[137],"soft":[138],"log":[139],"transformation":[140],"samples.":[143],"We":[144],"evaluate":[145,199],"these":[146],"other":[148],"previously":[149],"proposed":[150],"modification":[151],"several":[153,179,200],"challenging":[154],"target":[155],"distributions,":[156],"including":[157],"horseshoe":[160],"logistic":[161],"regression":[162],"model.":[163],"Our":[164],"experiments":[165],"our":[169],"modifications,":[170],"NVPs":[175,214],"posteriors":[177],"thousand":[180],"dimensions":[181],"heavy":[183],"tails":[184],"possible,":[186],"allowing":[187],"marginal":[191],"likelihood":[192],"estimation":[193],"via":[194],"importance":[195],"sampling.":[196],"Moreover,":[197],"common":[201],"techniques":[203],"architecture":[205],"choices":[206],"provide":[208,222],"practical":[209],"advise":[210],"variational":[217],"inference.":[218],"Finally,":[219],"also":[221],"new":[223],"theoretical":[226],"justification":[227],"optimizing":[229],"evidence":[231],"lower":[232],"bound":[233],"leads":[237],"distribution":[241],"coverage.":[242]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-06-13T06:13:01.061226","created_date":"2025-10-10T00:00:00"}