{"id":"https://openalex.org/W4416195015","doi":"https://doi.org/10.1145/3768292.3770442","title":"Discrete Flow Matching is a Surprisingly Effective Post-training Method to Address Compound Error in Autoregressive Models","display_name":"Discrete Flow Matching is a Surprisingly Effective Post-training Method to Address Compound Error in Autoregressive Models","publication_year":2025,"publication_date":"2025-11-14","ids":{"openalex":"https://openalex.org/W4416195015","doi":"https://doi.org/10.1145/3768292.3770442"},"language":null,"primary_location":{"id":"doi:10.1145/3768292.3770442","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3768292.3770442","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 6th ACM International Conference on AI in Finance","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3768292.3770442","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5001542402","display_name":"Kang Li","orcid":"https://orcid.org/0009-0007-7847-9292"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Kang Li","raw_affiliation_strings":["Department of Statistics, University of Oxford, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0007-7847-9292","affiliations":[{"raw_affiliation_string":"Department of Statistics, University of Oxford, Oxford, United Kingdom","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039641037","display_name":"Bidipta Sarkar","orcid":"https://orcid.org/0000-0002-0584-3504"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Bidipta Sarkar","raw_affiliation_strings":["Foerster Lab for AI Research, University of Oxford, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0002-0584-3504","affiliations":[{"raw_affiliation_string":"Foerster Lab for AI Research, University of Oxford, Oxford, United Kingdom","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112123269","display_name":"Xiong Zheng","orcid":"https://orcid.org/0000-0001-5707-5415"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Zheng Xiong","raw_affiliation_strings":["Department of Computer Science, University of Oxford, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0001-5707-5415","affiliations":[{"raw_affiliation_string":"Department of Computer Science, University of Oxford, Oxford, United Kingdom","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5061498010","display_name":"S. Frey","orcid":"https://orcid.org/0009-0001-6265-9607"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Sascha Frey","raw_affiliation_strings":["Department of Computer Science, University of Oxford, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0001-6265-9607","affiliations":[{"raw_affiliation_string":"Department of Computer Science, University of Oxford, Oxford, United Kingdom","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103326031","display_name":"Zilin Wang","orcid":"https://orcid.org/0009-0003-6062-3015"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Zilin Wang","raw_affiliation_strings":["Foerster Lab for AI Research, University of Oxford, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0003-6062-3015","affiliations":[{"raw_affiliation_string":"Foerster Lab for AI Research, University of Oxford, Oxford, United Kingdom","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5008328772","display_name":"Frensi Zejnullahu","orcid":"https://orcid.org/0009-0000-1947-6154"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Frensi Zejnullahu","raw_affiliation_strings":["Department of Statistics, University of Oxford, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0000-1947-6154","affiliations":[{"raw_affiliation_string":"Department of Statistics, University of Oxford, Oxford, United Kingdom","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5018185690","display_name":"Alfred Backhouse","orcid":"https://orcid.org/0009-0008-2218-1734"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Alfred Backhouse","raw_affiliation_strings":["Department of Computer Science, University of Oxford, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0008-2218-1734","affiliations":[{"raw_affiliation_string":"Department of Computer Science, University of Oxford, Oxford, United Kingdom","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090331439","display_name":"Stefan Zohren","orcid":"https://orcid.org/0000-0002-3392-0394"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]},{"id":"https://openalex.org/I4210146410","display_name":"Science Oxford","ror":"https://ror.org/04j8yhy50","country_code":"GB","type":"nonprofit","lineage":["https://openalex.org/I4210146410"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Stefan Zohren","raw_affiliation_strings":["Department of Engineering Science, University of Oxford, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0002-3392-0394","affiliations":[{"raw_affiliation_string":"Department of Engineering Science, University of Oxford, Oxford, United Kingdom","institution_ids":["https://openalex.org/I4210146410","https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5020492975","display_name":"Anisoara Calinescu","orcid":"https://orcid.org/0000-0003-2082-734X"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Anisoara Calinescu","raw_affiliation_strings":["Department of Computer Science, University of Oxford, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0003-2082-734X","affiliations":[{"raw_affiliation_string":"Department of Computer Science, University of Oxford, Oxford, United Kingdom","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039600886","display_name":"Mihai Cucuringu","orcid":"https://orcid.org/0000-0002-8464-2152"},"institutions":[{"id":"https://openalex.org/I4210098931","display_name":"Quantitative BioSciences","ror":"https://ror.org/010brsj79","country_code":"US","type":"nonprofit","lineage":["https://openalex.org/I4210098931"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mihai Cucuringu","raw_affiliation_strings":["Department of Mathematics, University of California Los Angeles, Los Angeles, USA and Department of Statistics & Oxford Man Institute of Quantitative Finance, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0002-8464-2152","affiliations":[{"raw_affiliation_string":"Department of Mathematics, University of California Los Angeles, Los Angeles, USA and Department of Statistics & Oxford Man Institute of Quantitative Finance, Oxford, United Kingdom","institution_ids":["https://openalex.org/I4210098931"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5059094093","display_name":"Jakob Foerster","orcid":"https://orcid.org/0000-0001-9688-2498"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Jakob Foerster","raw_affiliation_strings":["Foerster Lab for AI Research, University of Oxford, Oxford, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0001-9688-2498","affiliations":[{"raw_affiliation_string":"Foerster Lab for AI Research, University of Oxford, Oxford, United Kingdom","institution_ids":["https://openalex.org/I40120149"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":11,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.29872938,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"211","last_page":"219"},"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":0.7720999717712402,"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":0.7720999717712402,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.0430000014603138,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.03550000116229057,"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/autoregressive-model","display_name":"Autoregressive model","score":0.7239999771118164},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5400999784469604},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.5167999863624573},{"id":"https://openalex.org/keywords/design-for-manufacturability","display_name":"Design for manufacturability","score":0.4894999861717224},{"id":"https://openalex.org/keywords/flow","display_name":"Flow (mathematics)","score":0.4690000116825104},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.4106999933719635},{"id":"https://openalex.org/keywords/reduction","display_name":"Reduction (mathematics)","score":0.40540000796318054},{"id":"https://openalex.org/keywords/point","display_name":"Point (geometry)","score":0.38909998536109924},{"id":"https://openalex.org/keywords/residual","display_name":"Residual","score":0.38839998841285706}],"concepts":[{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.7239999771118164},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.629800021648407},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6227999925613403},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5400999784469604},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.5167999863624573},{"id":"https://openalex.org/C62064638","wikidata":"https://www.wikidata.org/wiki/Q553878","display_name":"Design for manufacturability","level":2,"score":0.4894999861717224},{"id":"https://openalex.org/C38349280","wikidata":"https://www.wikidata.org/wiki/Q1434290","display_name":"Flow (mathematics)","level":2,"score":0.4690000116825104},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.4106999933719635},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.40540000796318054},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.38909998536109924},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.38839998841285706},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.3817000091075897},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.3628999888896942},{"id":"https://openalex.org/C48677424","wikidata":"https://www.wikidata.org/wiki/Q6888088","display_name":"Mode (computer interface)","level":2,"score":0.36230000853538513},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.35920000076293945},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.352400004863739},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.3472000062465668},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3411000072956085},{"id":"https://openalex.org/C45493050","wikidata":"https://www.wikidata.org/wiki/Q7884934","display_name":"Unified Model","level":2,"score":0.3221000134944916},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.31619998812675476},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3073999881744385},{"id":"https://openalex.org/C183322885","wikidata":"https://www.wikidata.org/wiki/Q17007702","display_name":"Context model","level":3,"score":0.2921999990940094},{"id":"https://openalex.org/C149441793","wikidata":"https://www.wikidata.org/wiki/Q200726","display_name":"Probability distribution","level":2,"score":0.29030001163482666},{"id":"https://openalex.org/C55689738","wikidata":"https://www.wikidata.org/wiki/Q15963867","display_name":"Discrete time and continuous time","level":2,"score":0.27900001406669617},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.27079999446868896},{"id":"https://openalex.org/C103088060","wikidata":"https://www.wikidata.org/wiki/Q1062839","display_name":"Error detection and correction","level":2,"score":0.2660999894142151},{"id":"https://openalex.org/C159694833","wikidata":"https://www.wikidata.org/wiki/Q2321565","display_name":"Iterative method","level":2,"score":0.25290000438690186},{"id":"https://openalex.org/C167085575","wikidata":"https://www.wikidata.org/wiki/Q6803654","display_name":"Mean squared prediction error","level":2,"score":0.2524999976158142}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3768292.3770442","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3768292.3770442","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 6th ACM International Conference on AI in Finance","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3768292.3770442","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3768292.3770442","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 6th ACM International Conference on AI in Finance","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":3,"referenced_works":["https://openalex.org/W2988975212","https://openalex.org/W4313021454","https://openalex.org/W4388994542"],"related_works":[],"abstract_inverted_index":{"Autoregressive":[0],"(AR)":[1],"models":[2,40,83,127],"suffer":[3,88],"from":[4,89],"compound":[5,123],"error":[6,52,144],"accumulation,":[7],"where":[8,109],"prediction":[9],"errors":[10],"propagate":[11],"and":[12,50],"amplify":[13],"throughout":[14,135,146],"sequence":[15],"generation.":[16],"We":[17],"introduce":[18],"unified":[19,93,119,138],"discrete":[20,85,99],"flow":[21,70,94],"matching":[22,95],"(DFM)":[23],"as":[24,65],"a":[25,129,141,154],"post-training":[26,151],"framework":[27,59,96],"that":[28,87,118,132],"eliminates":[29,122],"this":[30],"fundamental":[31],"limitation":[32],"of":[33],"AR":[34,39,126],"models.":[35],"By":[36],"converting":[37],"pretrained":[38,62],"into":[41],"flow-based":[42],"generators,":[43],"our":[44,92,115],"approach":[45,152],"enables":[46],"bidirectional":[47],"context":[48],"integration":[49],"iterative":[51],"correction":[53],"while":[54,125],"preserving":[55],"learned":[56],"representations.":[57],"Our":[58],"leverages":[60],"the":[61,66,73,147],"model\u2019s":[63],"distribution":[64],"starting":[67],"point":[68],"for":[69],"matching,":[71],"reducing":[72],"learning":[74],"burden":[75],"compared":[76],"to":[77],"noise-based":[78],"approaches.":[79],"Unlike":[80],"BERT-style":[81],"masked":[82],"with":[84],"diffusion":[86],"mode":[90,106],"averaging,":[91],"successfully":[97],"learns":[98],"patterns,":[100],"achieving":[101],"99.9%":[102],"validity":[103],"on":[104],"solving":[105],"averaging":[107],"tasks":[108],"traditional":[110],"methods":[111],"completely":[112,121],"fail.":[113],"Critically,":[114],"experiments":[116],"demonstrate":[117],"DFM":[120,139],"error:":[124],"show":[128],"Wasserstein":[130,158],"distance":[131],"is":[133],"increasing":[134],"generation":[136],"progress,":[137],"maintains":[140],"nearly":[142],"flat":[143],"profile":[145],"entire":[148],"sequence.":[149],"This":[150],"achieves":[153],"48.6%":[155],"reduction":[156],"in":[157],"distance.":[159]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-11-14T00:00:00"}