{"id":"https://openalex.org/W4392903956","doi":"https://doi.org/10.1109/icassp48485.2024.10447605","title":"Prompting Large Language Models with Speech Recognition Abilities","display_name":"Prompting Large Language Models with Speech Recognition Abilities","publication_year":2024,"publication_date":"2024-03-18","ids":{"openalex":"https://openalex.org/W4392903956","doi":"https://doi.org/10.1109/icassp48485.2024.10447605"},"language":"en","primary_location":{"id":"doi:10.1109/icassp48485.2024.10447605","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp48485.2024.10447605","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","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/A5006814826","display_name":"Yassir Fathullah","orcid":null},"institutions":[{"id":"https://openalex.org/I241749","display_name":"University of Cambridge","ror":"https://ror.org/013meh722","country_code":"GB","type":"education","lineage":["https://openalex.org/I241749"]}],"countries":["GB"],"is_corresponding":true,"raw_author_name":"Yassir Fathullah","raw_affiliation_strings":["Meta AI","University of Cambridge"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]},{"raw_affiliation_string":"University of Cambridge","institution_ids":["https://openalex.org/I241749"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103012144","display_name":"Chunyang Wu","orcid":"https://orcid.org/0000-0002-5796-8288"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chunyang Wu","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5045428440","display_name":"Egor Lakomkin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Egor Lakomkin","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113970008","display_name":"Junteng Jia","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Junteng Jia","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047358828","display_name":"Yuan Shangguan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuan Shangguan","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100343462","display_name":"Ke Li","orcid":"https://orcid.org/0000-0001-9206-0892"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ke Li","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103232491","display_name":"Jinxi Guo","orcid":"https://orcid.org/0000-0001-9563-7351"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jinxi Guo","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110635444","display_name":"Wenhan Xiong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wenhan Xiong","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074237839","display_name":"Jay Mahadeokar","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jay Mahadeokar","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066166549","display_name":"Ozlem Kalinli","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ozlem Kalinli","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047073253","display_name":"Christian Fuegen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Christian Fuegen","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5113773386","display_name":"Mike Seltzer","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mike Seltzer","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":12,"corresponding_author_ids":["https://openalex.org/A5006814826"],"corresponding_institution_ids":["https://openalex.org/I241749"],"apc_list":null,"apc_paid":null,"fwci":28.7012,"has_fulltext":false,"cited_by_count":94,"citation_normalized_percentile":{"value":0.9976276,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":90,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"13351","last_page":"13355"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9998000264167786,"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":0.9998000264167786,"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":0.9997000098228455,"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":0.9997000098228455,"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/computer-science","display_name":"Computer science","score":0.7923476696014404},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.7024349570274353},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.6999742984771729},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.5939632654190063},{"id":"https://openalex.org/keywords/automatic-summarization","display_name":"Automatic summarization","score":0.583234429359436},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5670614242553711},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.49051302671432495},{"id":"https://openalex.org/keywords/acoustic-model","display_name":"Acoustic model","score":0.4809079170227051},{"id":"https://openalex.org/keywords/audio-mining","display_name":"Audio mining","score":0.41153115034103394},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4109063446521759},{"id":"https://openalex.org/keywords/speech-processing","display_name":"Speech processing","score":0.3232764005661011}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7923476696014404},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.7024349570274353},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.6999742984771729},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.5939632654190063},{"id":"https://openalex.org/C170858558","wikidata":"https://www.wikidata.org/wiki/Q1394144","display_name":"Automatic summarization","level":2,"score":0.583234429359436},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5670614242553711},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.49051302671432495},{"id":"https://openalex.org/C155635449","wikidata":"https://www.wikidata.org/wiki/Q4674699","display_name":"Acoustic model","level":3,"score":0.4809079170227051},{"id":"https://openalex.org/C157968479","wikidata":"https://www.wikidata.org/wiki/Q3079876","display_name":"Audio mining","level":4,"score":0.41153115034103394},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4109063446521759},{"id":"https://openalex.org/C61328038","wikidata":"https://www.wikidata.org/wiki/Q3358061","display_name":"Speech processing","level":2,"score":0.3232764005661011},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.0},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp48485.2024.10447605","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp48485.2024.10447605","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":35,"referenced_works":["https://openalex.org/W1583837637","https://openalex.org/W2127141656","https://openalex.org/W2914304175","https://openalex.org/W2963211188","https://openalex.org/W2973049979","https://openalex.org/W3095410713","https://openalex.org/W3097777922","https://openalex.org/W3198429080","https://openalex.org/W3212799896","https://openalex.org/W3213029956","https://openalex.org/W4224308101","https://openalex.org/W4226033575","https://openalex.org/W4292779060","https://openalex.org/W4308760226","https://openalex.org/W4311642023","https://openalex.org/W4320722432","https://openalex.org/W4322718191","https://openalex.org/W4323572061","https://openalex.org/W4366850747","https://openalex.org/W4372260518","https://openalex.org/W4377130946","https://openalex.org/W6738045163","https://openalex.org/W6759579507","https://openalex.org/W6778883912","https://openalex.org/W6779919476","https://openalex.org/W6780218876","https://openalex.org/W6796581206","https://openalex.org/W6804030475","https://openalex.org/W6810081322","https://openalex.org/W6810673746","https://openalex.org/W6849928207","https://openalex.org/W6850503672","https://openalex.org/W6850625674","https://openalex.org/W6851950068","https://openalex.org/W6853249747"],"related_works":["https://openalex.org/W2157598242","https://openalex.org/W2126322296","https://openalex.org/W3159882232","https://openalex.org/W4241650944","https://openalex.org/W2620660273","https://openalex.org/W4243125559","https://openalex.org/W2964829415","https://openalex.org/W2131711534","https://openalex.org/W642007152","https://openalex.org/W2341426843"],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2],"(LLMs)":[3],"have":[4],"proven":[5],"themselves":[6],"highly":[7],"flexible,":[8],"able":[9],"to":[10,44,56,66,103,130,141,157,197],"solve":[11],"a":[12,38,51,93],"wide":[13],"range":[14],"of":[15,33,53,181],"generative":[16],"tasks,":[17],"such":[18],"as":[19,81],"abstractive":[20],"summarization":[21],"and":[22,73,112,151],"open-ended":[23],"question":[24],"answering.":[25],"In":[26],"this":[27],"paper":[28],"we":[29,126],"extend":[30],"the":[31,57,61,77,97,133,148,153,174,188,193],"capabilities":[32],"LLM":[34,62,134,175],"by":[35,107],"directly":[36,49],"attaching":[37],"small":[39],"audio":[40,54,149,154,189],"encoder":[41,95,155,190],"allowing":[42],"it":[43,102],"perform":[45,113,127],"speech":[46,69,115],"recognition.":[47],"By":[48],"prepending":[50],"sequence":[52],"embeddings":[55],"text":[58],"token":[59],"embeddings,":[60],"can":[63,135],"be":[64,74,136],"converted":[65],"an":[67],"automatic":[68],"recognition":[70],"(ASR)":[71],"system,":[72],"used":[75,186],"in":[76,110,187],"exact":[78],"same":[79],"manner":[80],"its":[82,143],"textual":[83],"counterpart.":[84],"Experiments":[85],"on":[86,122,199],"Multilingual":[87],"LibriSpeech":[88],"(MLS)":[89],"show":[90,166],"that":[91,167],"incorporating":[92],"conformer":[94],"into":[96],"open":[98],"sourced":[99],"LLaMA-7B":[100],"allows":[101],"outperform":[104],"monolingual":[105],"baselines":[106],"18%":[108],"relatively":[109],"WER":[111],"multilingual":[114,168],"recognition,":[116],"despite":[117],"LLaMA":[118],"being":[119],"trained":[120],"overwhelmingly":[121],"English":[123],"text.":[124],"Furthermore,":[125],"ablation":[128],"studies":[129,165],"investigate":[131],"whether":[132],"completely":[137],"frozen":[138],"during":[139],"training":[140],"maintain":[142],"original":[144],"capabilities,":[145],"scaling":[146],"up":[147,192],"encoder,":[150],"increasing":[152],"striding":[156],"generate":[158],"fewer":[159],"embeddings.":[160],"The":[161],"results":[162],"from":[163],"these":[164],"ASR":[169],"is":[170,176],"possible":[171],"even":[172],"when":[173,179],"frozen,":[177],"or":[178],"strides":[180],"almost":[182],"1":[183],"second":[184],"are":[185],"opening":[191],"possibility":[194],"for":[195],"LLMs":[196],"operate":[198],"long-form":[200],"audio.":[201]},"counts_by_year":[{"year":2026,"cited_by_count":11},{"year":2025,"cited_by_count":54},{"year":2024,"cited_by_count":28},{"year":2023,"cited_by_count":1}],"updated_date":"2026-06-05T09:01:59.212387","created_date":"2025-10-10T00:00:00"}