{"id":"https://openalex.org/W4383604245","doi":"https://doi.org/10.48550/arxiv.2307.01715","title":"Align With Purpose: Optimize Desired Properties in CTC Models with a General Plug-and-Play Framework","display_name":"Align With Purpose: Optimize Desired Properties in CTC Models with a General Plug-and-Play Framework","publication_year":2023,"publication_date":"2023-07-04","ids":{"openalex":"https://openalex.org/W4383604245","doi":"https://doi.org/10.48550/arxiv.2307.01715"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2307.01715","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2307.01715","pdf_url":"https://arxiv.org/pdf/2307.01715","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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","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/2307.01715","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5092427158","display_name":"Eliya Segev","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Segev, Eliya","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092427159","display_name":"Maya Alroy","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Alroy, Maya","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069912489","display_name":"Ronen Katsir","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Katsir, Ronen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5081397200","display_name":"Noam Wies","orcid":"https://orcid.org/0000-0002-1337-2298","observed_orcids":["https://orcid.org/0000-0002-1337-2298"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wies, Noam","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104140845","display_name":"Ayana Shenhav","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shenhav, Ayana","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092427161","display_name":"Yael Ben-Oren","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ben-Oren, Yael","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073388054","display_name":"David Zar","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zar, David","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5055559377","display_name":"Oren Tadmor","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tadmor, Oren","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068306589","display_name":"Jacob Bitterman","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bitterman, Jacob","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5075004917","display_name":"Amnon Shashua","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shashua, Amnon","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5088431112","display_name":"Tal Rosenwein","orcid":"https://orcid.org/0009-0006-8127-4354","observed_orcids":["https://orcid.org/0009-0006-8127-4354"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rosenwein, Tal","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":4,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9990000128746033,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9990000128746033,"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/T10860","display_name":"Speech and Audio Processing","score":0.00019999999494757503,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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/T10028","display_name":"Topic Modeling","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/connectionist-temporal-classification","display_name":"connectionist temporal classification","score":0.9860000014305115},{"id":"https://openalex.org/keywords/plug-and-play-framework","display_name":"plug-and-play framework","score":0.7829999923706055},{"id":"https://openalex.org/keywords/sequence-to-sequence-models","display_name":"sequence-to-sequence models","score":0.7609999775886536},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"speech recognition","score":0.8700000047683716},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"word error rate","score":0.9089999794960022},{"id":"https://openalex.org/keywords/alignment-loss","display_name":"alignment loss","score":0.35899999737739563},{"id":"https://openalex.org/keywords/latency-optimization","display_name":"latency optimization","score":0.4180000126361847}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6888289451599121},{"id":"https://openalex.org/C2780767217","wikidata":"https://www.wikidata.org/wiki/Q5532421","display_name":"Generality","level":2,"score":0.62831050157547},{"id":"https://openalex.org/C189950617","wikidata":"https://www.wikidata.org/wiki/Q937228","display_name":"Property (philosophy)","level":2,"score":0.5254877209663391},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.48227575421333313},{"id":"https://openalex.org/C2780310539","wikidata":"https://www.wikidata.org/wiki/Q12547192","display_name":"Imperfect","level":2,"score":0.4625697731971741},{"id":"https://openalex.org/C8521452","wikidata":"https://www.wikidata.org/wiki/Q203790","display_name":"Connectionism","level":3,"score":0.4624066650867462},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.45922887325286865},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.45433124899864197},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.4370821416378021},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.43530362844467163},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.42043042182922363},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3530612885951996},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3413805067539215},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.19865939021110535},{"id":"https://openalex.org/C94375191","wikidata":"https://www.wikidata.org/wiki/Q11205","display_name":"Arithmetic","level":1,"score":0.15696892142295837},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.15350019931793213},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","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/C542102704","wikidata":"https://www.wikidata.org/wiki/Q183257","display_name":"Psychotherapist","level":1,"score":0.0},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C54355233","wikidata":"https://www.wikidata.org/wiki/Q7162","display_name":"Genetics","level":1,"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/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","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}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2307.01715","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2307.01715","pdf_url":"https://arxiv.org/pdf/2307.01715","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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2307.01715","doi":"https://doi.org/10.48550/arxiv.2307.01715","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2307.01715","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:2307.01715","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2307.01715","pdf_url":"https://arxiv.org/pdf/2307.01715","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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4383604245.pdf","grobid_xml":"https://content.openalex.org/works/W4383604245.grobid-xml"},"referenced_works_count":16,"referenced_works":["https://openalex.org/W1828163288","https://openalex.org/W1922655562","https://openalex.org/W2520160253","https://openalex.org/W2547875792","https://openalex.org/W2936774411","https://openalex.org/W2964121744","https://openalex.org/W2968917279","https://openalex.org/W3001899777","https://openalex.org/W3097777922","https://openalex.org/W3167895882","https://openalex.org/W4225934689","https://openalex.org/W4296069356","https://openalex.org/W4306705025","https://openalex.org/W4385822293","https://openalex.org/W4385823471","https://openalex.org/W4394642489"],"related_works":["https://openalex.org/W7117397260","https://openalex.org/W4407183551","https://openalex.org/W2395416438","https://openalex.org/W4306705025","https://openalex.org/W3208871315","https://openalex.org/W4297841727","https://openalex.org/W4392902865","https://openalex.org/W3023277061","https://openalex.org/W3016120074","https://openalex.org/W2892209435"],"abstract_inverted_index":{"Connectionist":[0],"Temporal":[1],"Classification":[2],"(CTC)":[3],"is":[4],"a":[5,72,78,105,124,205,216,241,257],"widely":[6],"used":[7],"criterion":[8],"for":[9,76,211],"training":[10,163],"supervised":[11],"sequence-to-sequence":[12],"(seq2seq)":[13],"models.":[14,225],"It":[15],"enables":[16,120],"learning":[17],"the":[18,34,38,85,93,116,142,171,191,212,223,227],"relations":[19],"between":[20,131],"input":[21],"and":[22,48,128,134,149,160,185,210,262,271],"output":[23],"sequences,":[24],"termed":[25],"alignments,":[26],"by":[27,91],"marginalizing":[28],"over":[29,222],"perfect":[30,47,133],"alignments":[31,50,102],"(that":[32],"yield":[33],"ground":[35],"truth),":[36],"at":[37],"expense":[39],"of":[40,46,53,123,126,144,155,162,173,197,219,229,243,260],"imperfect":[41,49,135],"alignments.":[42,136],"This":[43],"binary":[44],"differentiation":[45,130],"falls":[51],"short":[52],"capturing":[54],"other":[55,63,267,274],"essential":[56],"alignment":[57],"properties":[58],"that":[59,90,100],"hold":[60],"significance":[61],"in":[62,81,115,141,153,201,208],"real-world":[64],"applications.":[65],"Here":[66],"we":[67,176,193,214],"propose":[68],"$\\textit{Align":[69],"With":[70],"Purpose}$,":[71],"$\\textbf{general":[73],"Plug-and-Play":[74],"framework}$":[75],"enhancing":[77],"desired":[79,106],"property":[80,156],"models":[82],"trained":[83],"with":[84,95,204],"CTC":[86,94,117],"criterion.":[87],"We":[88,137],"do":[89],"complementing":[92],"an":[96,195],"additional":[97],"loss":[98,118,269],"term":[99],"prioritizes":[101],"according":[103],"to":[104,166,179,199,238,266,272],"property.":[107],"Our":[108],"method":[109,251],"does":[110],"not":[111],"require":[112],"any":[113],"intervention":[114],"function,":[119],"easy":[121],"optimization":[122,203],"variety":[125],"properties,":[127],"allows":[129],"both":[132],"apply":[138,177],"our":[139,174,230,250],"framework":[140],"domain":[143],"Automatic":[145],"Speech":[146],"Recognition":[147],"(ASR)":[148],"show":[150],"its":[151],"generality":[152],"terms":[154],"selection,":[157],"architectural":[158],"choice,":[159],"scale":[161,242],"dataset":[164],"(up":[165],"280,000":[167],"hours).":[168],"To":[169,226],"demonstrate":[170],"effectiveness":[172],"framework,":[175],"it":[178],"two":[180],"unrelated":[181],"properties:":[182],"emission":[183],"time":[184],"word":[186],"error":[187],"rate":[188],"(WER).":[189],"For":[190],"former,":[192],"report":[194,215],"improvement":[196,218],"up":[198],"570ms":[200],"latency":[202],"minor":[206],"reduction":[207],"WER,":[209],"latter,":[213],"relative":[217],"4.5%":[220],"WER":[221],"baseline":[224],"best":[228],"knowledge,":[231],"these":[232],"applications":[233],"have":[234],"never":[235],"been":[236],"demonstrated":[237],"work":[239],"on":[240],"data":[244],"as":[245,247],"large":[246],"ours.":[248],"Notably,":[249],"can":[252,263],"be":[253,264],"implemented":[254],"using":[255],"only":[256],"few":[258],"lines":[259],"code,":[261],"extended":[265],"alignment-free":[268],"functions":[270],"domains":[273],"than":[275],"ASR.":[276]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":2}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2023-07-08T00:00:00"}