{"id":"https://openalex.org/W7148462517","doi":"https://doi.org/10.1109/asru65441.2025.11434766","title":"Towards Efficient Speech-Text Jointly Decoding within One Speech Language Model","display_name":"Towards Efficient Speech-Text Jointly Decoding within One Speech Language Model","publication_year":2025,"publication_date":"2025-12-06","ids":{"openalex":"https://openalex.org/W7148462517","doi":"https://doi.org/10.1109/asru65441.2025.11434766"},"language":null,"primary_location":{"id":"doi:10.1109/asru65441.2025.11434766","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asru65441.2025.11434766","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","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":null,"display_name":"Haibin Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Haibin Wu","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yuxuan Hu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuxuan Hu","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Ruchao Fan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ruchao Fan","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Xiaofei Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiaofei Wang","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Kenichi Kumatani","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kenichi Kumatani","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Bo Ren","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bo Ren","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Jianwei Yu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jianwei Yu","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Heng Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Heng Lu","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Lijuan Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lijuan Wang","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yao Qian","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yao Qian","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"Jinyu Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jinyu Li","raw_affiliation_strings":["Microsoft,USA"],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"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":0,"citation_normalized_percentile":{"value":0.87619416,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"7"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.5144000053405762,"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/T10028","display_name":"Topic Modeling","score":0.5144000053405762,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.32420000433921814,"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/T12031","display_name":"Speech and dialogue systems","score":0.05009999871253967,"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/decoding-methods","display_name":"Decoding methods","score":0.8853999972343445},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.8547999858856201},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.691100001335144},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5867999792098999},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.5612999796867371},{"id":"https://openalex.org/keywords/joint","display_name":"Joint (building)","score":0.5113000273704529}],"concepts":[{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.8853999972343445},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.8547999858856201},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8187000155448914},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.691100001335144},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5924999713897705},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5867999792098999},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.5612999796867371},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5259000062942505},{"id":"https://openalex.org/C18555067","wikidata":"https://www.wikidata.org/wiki/Q8375051","display_name":"Joint (building)","level":2,"score":0.5113000273704529},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5052000284194946},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.4339999854564667},{"id":"https://openalex.org/C2776230583","wikidata":"https://www.wikidata.org/wiki/Q1322198","display_name":"Spoken language","level":2,"score":0.38999998569488525},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.3684000074863434},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.29030001163482666},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.29030001163482666},{"id":"https://openalex.org/C35639132","wikidata":"https://www.wikidata.org/wiki/Q7452468","display_name":"Sequence labeling","level":3,"score":0.2775999903678894}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/asru65441.2025.11434766","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asru65441.2025.11434766","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","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":6,"referenced_works":["https://openalex.org/W2963339397","https://openalex.org/W4402670057","https://openalex.org/W4406461437","https://openalex.org/W4412888872","https://openalex.org/W4412944968","https://openalex.org/W4417305271"],"related_works":[],"abstract_inverted_index":{"Speech":[0],"language":[1,63],"models":[2],"(Speech":[3],"LMs)":[4],"enable":[5],"end-to-end":[6],"speech-text":[7,24,46],"modeling":[8],"within":[9],"a":[10,14,29,55,98],"single":[11],"model,":[12,64],"offering":[13],"promising":[15],"direction":[16],"for":[17],"spoken":[18],"dialogue":[19],"systems.":[20],"The":[21],"choice":[22],"of":[23],"jointly":[25],"decoding":[26,47,109],"paradigm":[27],"plays":[28],"critical":[30],"role":[31],"in":[32],"performance,":[33],"efficiency,":[34],"and":[35,51,67],"alignment":[36],"quality.":[37],"In":[38],"this":[39],"work,":[40],"we":[41,96,117],"systematically":[42],"compare":[43],"representative":[44],"joint":[45],"strategies\u2014including":[48],"the":[49,60,74,78],"interleaved,":[50],"parallel":[52],"generation":[53],"paradigms\u2014under":[54],"controlled":[56],"experimental":[57],"setup":[58],"using":[59],"same":[61],"base":[62],"speech":[65,127],"tokenizer":[66],"training":[68],"data.":[69],"Our":[70],"results":[71],"show":[72],"that":[73,104],"interleaved":[75,101],"approach":[76],"achieves":[77],"best":[79],"alignment.":[80],"However":[81],"it":[82],"suffers":[83],"from":[84],"slow":[85],"inference":[86],"due":[87],"to":[88,124],"long":[89],"token":[90],"sequence":[91],"length.":[92],"To":[93],"address":[94],"this,":[95],"propose":[97],"novel":[99],"early-stop":[100],"(ESI)":[102],"pattern":[103],"not":[105],"only":[106],"significantly":[107],"accelerates":[108],"but":[110],"also":[111],"yields":[112],"slightly":[113],"better":[114],"performance.":[115,129],"Additionally,":[116],"curate":[118],"high-quality":[119],"question":[120],"answering":[121],"(QA)":[122],"datasets":[123],"further":[125],"improve":[126],"QA":[128]},"counts_by_year":[],"updated_date":"2026-04-03T22:45:19.894376","created_date":"2026-02-14T00:00:00"}