{"id":"https://openalex.org/W7148575771","doi":"https://doi.org/10.1109/asru65441.2025.11434774","title":"Efficient Scaling for LLM-based ASR","display_name":"Efficient Scaling for LLM-based ASR","publication_year":2025,"publication_date":"2025-12-06","ids":{"openalex":"https://openalex.org/W7148575771","doi":"https://doi.org/10.1109/asru65441.2025.11434774"},"language":null,"primary_location":{"id":"doi:10.1109/asru65441.2025.11434774","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asru65441.2025.11434774","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":"https://openalex.org/A5132756133","display_name":"Bingshen Mu","orcid":null},"institutions":[{"id":"https://openalex.org/I2250653659","display_name":"Tencent (China)","ror":"https://ror.org/00hhjss72","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250653659"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Bingshen Mu","raw_affiliation_strings":["Tencent AI Lab"],"affiliations":[{"raw_affiliation_string":"Tencent AI Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130089581","display_name":"Yiwen Shao","orcid":null},"institutions":[{"id":"https://openalex.org/I2250653659","display_name":"Tencent (China)","ror":"https://ror.org/00hhjss72","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250653659"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yiwen Shao","raw_affiliation_strings":["Tencent AI Lab"],"affiliations":[{"raw_affiliation_string":"Tencent AI Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100649112","display_name":"Kun Wei","orcid":null},"institutions":[{"id":"https://openalex.org/I2250653659","display_name":"Tencent (China)","ror":"https://ror.org/00hhjss72","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250653659"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Kun Wei","raw_affiliation_strings":["Tencent AI Lab"],"affiliations":[{"raw_affiliation_string":"Tencent AI Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132812568","display_name":"Dong Yu","orcid":null},"institutions":[{"id":"https://openalex.org/I2250653659","display_name":"Tencent (China)","ror":"https://ror.org/00hhjss72","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250653659"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dong Yu","raw_affiliation_strings":["Tencent AI Lab"],"affiliations":[{"raw_affiliation_string":"Tencent AI Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5132816141","display_name":"Lei Xie","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lei Xie","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5132756133"],"corresponding_institution_ids":["https://openalex.org/I2250653659"],"apc_list":null,"apc_paid":null,"fwci":4.3637,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.95468938,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":99},"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.1940000057220459,"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":0.1940000057220459,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.08110000193119049,"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/T12072","display_name":"Machine Learning and Algorithms","score":0.06729999929666519,"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/encoder","display_name":"Encoder","score":0.7419000267982483},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.7386000156402588},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.6001999974250793},{"id":"https://openalex.org/keywords/joint","display_name":"Joint (building)","score":0.4885999858379364},{"id":"https://openalex.org/keywords/work","display_name":"Work (physics)","score":0.3968999981880188},{"id":"https://openalex.org/keywords/encoding","display_name":"Encoding (memory)","score":0.38370001316070557},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.3513000011444092}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7511000037193298},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.7419000267982483},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.7386000156402588},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.6001999974250793},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5389000177383423},{"id":"https://openalex.org/C18555067","wikidata":"https://www.wikidata.org/wiki/Q8375051","display_name":"Joint (building)","level":2,"score":0.4885999858379364},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4327000081539154},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4235999882221222},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.3968999981880188},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.38370001316070557},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3513000011444092},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.3481999933719635},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.3082999885082245},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.30720001459121704},{"id":"https://openalex.org/C2778915421","wikidata":"https://www.wikidata.org/wiki/Q3643177","display_name":"Performance improvement","level":2,"score":0.28519999980926514},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2653999924659729},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.2599000036716461},{"id":"https://openalex.org/C61328038","wikidata":"https://www.wikidata.org/wiki/Q3358061","display_name":"Speech processing","level":2,"score":0.2502000033855438}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/asru65441.2025.11434774","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asru65441.2025.11434774","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":[{"display_name":"Affordable and clean energy","score":0.5172529220581055,"id":"https://metadata.un.org/sdg/7"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":23,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W2127141656","https://openalex.org/W2160815625","https://openalex.org/W2327501763","https://openalex.org/W2766219058","https://openalex.org/W2962826786","https://openalex.org/W2963211739","https://openalex.org/W2963242190","https://openalex.org/W3196117288","https://openalex.org/W3198028464","https://openalex.org/W3203407300","https://openalex.org/W3207834622","https://openalex.org/W4297727296","https://openalex.org/W4385245566","https://openalex.org/W4385823001","https://openalex.org/W4392903921","https://openalex.org/W4393178509","https://openalex.org/W4400903651","https://openalex.org/W4404782909","https://openalex.org/W4405709396","https://openalex.org/W4406461489","https://openalex.org/W4408345589","https://openalex.org/W6959349094"],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"model":[2],"(LLM)-based":[3],"automatic":[4],"speech":[5,38],"recognition":[6],"(ASR)":[7],"achieves":[8],"strong":[9],"performance":[10,26,86],"but":[11],"often":[12],"incurs":[13],"high":[14],"computational":[15],"costs.":[16],"This":[17],"work":[18],"investigates":[19],"how":[20],"to":[21,47,88],"obtain":[22],"the":[23,37,44,53],"best":[24],"LLM-ASR":[25,70,119],"efficiently.":[27],"Through":[28],"comprehensive":[29],"and":[30],"controlled":[31],"experiments,":[32],"we":[33,65,101],"find":[34],"that":[35,106],"pretraining":[36],"encoder":[39],"before":[40],"integrating":[41],"it":[42],"with":[43,92],"LLM":[45],"leads":[46],"significantly":[48,93],"better":[49,85],"scaling":[50,104],"efficiency":[51],"than":[52],"standard":[54],"practice":[55],"of":[56,59],"joint":[57],"post-training":[58],"LLM-ASR.":[60],"Based":[61],"on":[62],"this":[63],"insight,":[64],"propose":[66],"a":[67,103,112],"new":[68],"multi-stage":[69],"training":[71,79],"strategy,":[72],"EFIN:":[73],"Encoder":[74],"First":[75],"Integration.":[76],"Among":[77],"all":[78],"strategies":[80],"evaluated,":[81],"EFIN":[82],"consistently":[83],"delivers":[84],"(relative":[87],"21.1":[89],"%":[90],"CERR)":[91],"lower":[94],"computation":[95,113],"budgets":[96],"($49.9":[97],"\\%$":[98],"FLOPs).":[99],"Furthermore,":[100],"derive":[102],"law":[105],"approximates":[107],"ASR":[108],"error":[109],"rates":[110],"as":[111],"function,":[114],"providing":[115],"practical":[116],"guidance":[117],"for":[118],"scaling.":[120]},"counts_by_year":[{"year":2026,"cited_by_count":2}],"updated_date":"2026-04-23T09:07:50.710637","created_date":"2026-04-03T00:00:00"}