{"id":"https://openalex.org/W4296907727","doi":"https://doi.org/10.1088/2632-2153/ac9455","title":"Gradients should stay on path: better estimators of the reverse- and forward KL divergence for normalizing flows","display_name":"Gradients should stay on path: better estimators of the reverse- and forward KL divergence for normalizing flows","publication_year":2022,"publication_date":"2022-09-22","ids":{"openalex":"https://openalex.org/W4296907727","doi":"https://doi.org/10.1088/2632-2153/ac9455"},"language":"en","primary_location":{"id":"doi:10.1088/2632-2153/ac9455","is_oa":true,"landing_page_url":"https://doi.org/10.1088/2632-2153/ac9455","pdf_url":"https://iopscience.iop.org/article/10.1088/2632-2153/ac9455/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/ac9455/pdf","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5012899215","display_name":"Lorenz Vaitl","orcid":"https://orcid.org/0009-0009-3818-8911"},"institutions":[{"id":"https://openalex.org/I4577782","display_name":"Technische Universit\u00e4t Berlin","ror":"https://ror.org/03v4gjf40","country_code":"DE","type":"education","lineage":["https://openalex.org/I4577782"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Lorenz Vaitl","raw_affiliation_strings":["Department of Electrical Engineering & Computer Science, Technische Universit\u00e4t Berlin, Machine Learning Group, Berlin, Germany"],"affiliations":[{"raw_affiliation_string":"Department of Electrical Engineering & Computer Science, Technische Universit\u00e4t Berlin, Machine Learning Group, Berlin, Germany","institution_ids":["https://openalex.org/I4577782"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033457512","display_name":"Kim A. Nicoli","orcid":"https://orcid.org/0000-0001-5933-1822"},"institutions":[{"id":"https://openalex.org/I4577782","display_name":"Technische Universit\u00e4t Berlin","ror":"https://ror.org/03v4gjf40","country_code":"DE","type":"education","lineage":["https://openalex.org/I4577782"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Kim A Nicoli","raw_affiliation_strings":["BIFOLD\u2014Berlin Institute for the Foundations of Learning and Data, Technische Universit\u00e4t Berlin, Berlin, Germany","Department of Electrical Engineering & Computer Science, Technische Universit\u00e4t Berlin, Machine Learning Group, Berlin, Germany"],"affiliations":[{"raw_affiliation_string":"BIFOLD\u2014Berlin Institute for the Foundations of Learning and Data, Technische Universit\u00e4t Berlin, Berlin, Germany","institution_ids":["https://openalex.org/I4577782"]},{"raw_affiliation_string":"Department of Electrical Engineering & Computer Science, Technische Universit\u00e4t Berlin, Machine Learning Group, Berlin, Germany","institution_ids":["https://openalex.org/I4577782"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5034738514","display_name":"Shinichi Nakajima","orcid":"https://orcid.org/0000-0003-3970-4569"},"institutions":[{"id":"https://openalex.org/I4210126580","display_name":"RIKEN Center for Advanced Intelligence Project","ror":"https://ror.org/03ckxwf91","country_code":"JP","type":"facility","lineage":["https://openalex.org/I4210110652","https://openalex.org/I4210126580"]},{"id":"https://openalex.org/I4577782","display_name":"Technische Universit\u00e4t Berlin","ror":"https://ror.org/03v4gjf40","country_code":"DE","type":"education","lineage":["https://openalex.org/I4577782"]}],"countries":["DE","JP"],"is_corresponding":false,"raw_author_name":"Shinichi Nakajima","raw_affiliation_strings":["RIKEN Center for AIP, 103-0027 Tokyo, Chuo City, Japan","BIFOLD\u2014Berlin Institute for the Foundations of Learning and Data, Technische Universit\u00e4t Berlin, Berlin, Germany","Department of Electrical Engineering & Computer Science, Technische Universit\u00e4t Berlin, Machine Learning Group, Berlin, Germany"],"affiliations":[{"raw_affiliation_string":"RIKEN Center for AIP, 103-0027 Tokyo, Chuo City, Japan","institution_ids":["https://openalex.org/I4210126580"]},{"raw_affiliation_string":"BIFOLD\u2014Berlin Institute for the Foundations of Learning and Data, Technische Universit\u00e4t Berlin, Berlin, Germany","institution_ids":["https://openalex.org/I4577782"]},{"raw_affiliation_string":"Department of Electrical Engineering & Computer Science, Technische Universit\u00e4t Berlin, Machine Learning Group, Berlin, Germany","institution_ids":["https://openalex.org/I4577782"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5107747746","display_name":"Pan Kessel","orcid":"https://orcid.org/0009-0006-5951-7188"},"institutions":[{"id":"https://openalex.org/I4577782","display_name":"Technische Universit\u00e4t Berlin","ror":"https://ror.org/03v4gjf40","country_code":"DE","type":"education","lineage":["https://openalex.org/I4577782"]}],"countries":["DE"],"is_corresponding":true,"raw_author_name":"Pan Kessel","raw_affiliation_strings":["Department of Electrical Engineering & Computer Science, Technische Universit\u00e4t Berlin, Machine Learning Group, Berlin, Germany","BIFOLD\u2014Berlin Institute for the Foundations of Learning and Data, Technische Universit\u00e4t Berlin, Berlin, Germany"],"affiliations":[{"raw_affiliation_string":"Department of Electrical Engineering & Computer Science, Technische Universit\u00e4t Berlin, Machine Learning Group, Berlin, Germany","institution_ids":["https://openalex.org/I4577782"]},{"raw_affiliation_string":"BIFOLD\u2014Berlin Institute for the Foundations of Learning and Data, Technische Universit\u00e4t Berlin, Berlin, Germany","institution_ids":["https://openalex.org/I4577782"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5107747746"],"corresponding_institution_ids":["https://openalex.org/I4577782"],"apc_list":{"value":1600,"currency":"GBP","value_usd":1962},"apc_paid":{"value":1600,"currency":"GBP","value_usd":1962},"fwci":1.7491,"has_fulltext":false,"cited_by_count":14,"citation_normalized_percentile":{"value":0.84896777,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":97,"max":98},"biblio":{"volume":"3","issue":"4","first_page":"045006","last_page":"045006"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11206","display_name":"Model Reduction and Neural Networks","score":0.994700014591217,"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"}},"topics":[{"id":"https://openalex.org/T11206","display_name":"Model Reduction and Neural Networks","score":0.994700014591217,"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"}},{"id":"https://openalex.org/T10597","display_name":"Nuclear reactor physics and engineering","score":0.9891999959945679,"subfield":{"id":"https://openalex.org/subfields/2202","display_name":"Aerospace Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9868999719619751,"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/estimator","display_name":"Estimator","score":0.8067995309829712},{"id":"https://openalex.org/keywords/divergence","display_name":"Divergence (linguistics)","score":0.7292319536209106},{"id":"https://openalex.org/keywords/path","display_name":"Path (computing)","score":0.6739844679832458},{"id":"https://openalex.org/keywords/convergence","display_name":"Convergence (economics)","score":0.588513195514679},{"id":"https://openalex.org/keywords/variance","display_name":"Variance (accounting)","score":0.4960082471370697},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.4950977861881256},{"id":"https://openalex.org/keywords/applied-mathematics","display_name":"Applied mathematics","score":0.477750688791275},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.4636557996273041},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4225141704082489},{"id":"https://openalex.org/keywords/mathematical-optimization","display_name":"Mathematical optimization","score":0.4079791307449341},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.3396601378917694},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.3278563618659973},{"id":"https://openalex.org/keywords/statistics","display_name":"Statistics","score":0.23954561352729797},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.09042006731033325},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.0900392234325409}],"concepts":[{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.8067995309829712},{"id":"https://openalex.org/C207390915","wikidata":"https://www.wikidata.org/wiki/Q1230525","display_name":"Divergence (linguistics)","level":2,"score":0.7292319536209106},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.6739844679832458},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.588513195514679},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.4960082471370697},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.4950977861881256},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.477750688791275},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.4636557996273041},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4225141704082489},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.4079791307449341},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3396601378917694},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.3278563618659973},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.23954561352729797},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.09042006731033325},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.0900392234325409},{"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/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C121955636","wikidata":"https://www.wikidata.org/wiki/Q4116214","display_name":"Accounting","level":1,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1088/2632-2153/ac9455","is_oa":true,"landing_page_url":"https://doi.org/10.1088/2632-2153/ac9455","pdf_url":"https://iopscience.iop.org/article/10.1088/2632-2153/ac9455/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"}],"best_oa_location":{"id":"doi:10.1088/2632-2153/ac9455","is_oa":true,"landing_page_url":"https://doi.org/10.1088/2632-2153/ac9455","pdf_url":"https://iopscience.iop.org/article/10.1088/2632-2153/ac9455/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":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4296907727.pdf","grobid_xml":"https://content.openalex.org/works/W4296907727.grobid-xml"},"referenced_works_count":68,"referenced_works":["https://openalex.org/W1513873506","https://openalex.org/W1779483307","https://openalex.org/W1959608418","https://openalex.org/W2119717200","https://openalex.org/W2238910925","https://openalex.org/W2735102987","https://openalex.org/W2769810959","https://openalex.org/W2785443624","https://openalex.org/W2804766086","https://openalex.org/W2804802260","https://openalex.org/W2893674448","https://openalex.org/W2924794921","https://openalex.org/W2941443959","https://openalex.org/W2952264928","https://openalex.org/W2953194650","https://openalex.org/W2963936947","https://openalex.org/W2972246420","https://openalex.org/W2973674981","https://openalex.org/W2979652999","https://openalex.org/W3006106210","https://openalex.org/W3007848206","https://openalex.org/W3012274818","https://openalex.org/W3034727712","https://openalex.org/W3042241440","https://openalex.org/W3048457240","https://openalex.org/W3089052316","https://openalex.org/W3161230803","https://openalex.org/W3167003437","https://openalex.org/W3181973887","https://openalex.org/W3212685810","https://openalex.org/W3217363338","https://openalex.org/W4283316023","https://openalex.org/W4283331950","https://openalex.org/W4286939909","https://openalex.org/W4287330004","https://openalex.org/W4287367937","https://openalex.org/W4287636448","https://openalex.org/W4287755451","https://openalex.org/W4289420068","https://openalex.org/W4290064682","https://openalex.org/W4294619240","https://openalex.org/W4297692234","https://openalex.org/W6637161497","https://openalex.org/W6638116569","https://openalex.org/W6640963894","https://openalex.org/W6677929280","https://openalex.org/W6690165720","https://openalex.org/W6728917834","https://openalex.org/W6739922093","https://openalex.org/W6742563595","https://openalex.org/W6747982189","https://openalex.org/W6751300517","https://openalex.org/W6752110675","https://openalex.org/W6755750391","https://openalex.org/W6760866730","https://openalex.org/W6767596605","https://openalex.org/W6779585612","https://openalex.org/W6779757826","https://openalex.org/W6783333127","https://openalex.org/W6783983540","https://openalex.org/W6785198518","https://openalex.org/W6789125565","https://openalex.org/W6790965610","https://openalex.org/W6793565200","https://openalex.org/W6798823188","https://openalex.org/W6838914865","https://openalex.org/W6839557214","https://openalex.org/W6841327968"],"related_works":["https://openalex.org/W4287880334","https://openalex.org/W4366700029","https://openalex.org/W3175363083","https://openalex.org/W4285230481","https://openalex.org/W3208095355","https://openalex.org/W4385769873","https://openalex.org/W3122602933","https://openalex.org/W2950038056","https://openalex.org/W1544940847","https://openalex.org/W2289285490"],"abstract_inverted_index":{"Abstract":[0],"We":[1,56],"show":[2],"how":[3],"to":[4,29,38,45,51,65,83],"use":[5],"the":[6,12,79],"path-wise":[7],"derivative":[8],"estimator":[9],"for":[10,17,87],"both":[11],"forward":[13],"reverse":[14],"Kullback\u2013Leibler":[15],"divergence":[16],"any":[18],"practically":[19],"invertible":[20],"normalizing":[21,85],"flow.":[22],"The":[23],"resulting":[24],"path-gradient":[25,60,75],"estimators":[26,76],"are":[27],"straightforward":[28],"implement,":[30],"have":[31],"lower":[32],"variance,":[33],"and":[34],"lead":[35],"not":[36],"only":[37],"faster":[39],"convergence":[40],"of":[41,69],"training":[42,61],"but":[43],"also":[44,57],"better":[46],"overall":[47],"approximation":[48],"results":[49],"compared":[50],"standard":[52,81],"total":[53],"gradient":[54],"estimators.":[55],"demonstrate":[58],"that":[59,74],"is":[62],"less":[63],"susceptible":[64],"mode-collapse.":[66],"In":[67],"light":[68],"our":[70],"results,":[71],"we":[72],"expect":[73],"will":[77],"become":[78],"new":[80],"method":[82],"train":[84],"flows":[86],"variational":[88],"inference.":[89]},"counts_by_year":[{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":6},{"year":2023,"cited_by_count":4}],"updated_date":"2026-03-27T14:29:43.386196","created_date":"2025-10-10T00:00:00"}