{"id":"https://openalex.org/W4416799130","doi":"https://doi.org/10.1109/apsipaasc65261.2025.11249262","title":"LLM-Driven Hypothesis Set Refinement for Enhanced ASR Post-Processing","display_name":"LLM-Driven Hypothesis Set Refinement for Enhanced ASR Post-Processing","publication_year":2025,"publication_date":"2025-10-22","ids":{"openalex":"https://openalex.org/W4416799130","doi":"https://doi.org/10.1109/apsipaasc65261.2025.11249262"},"language":null,"primary_location":{"id":"doi:10.1109/apsipaasc65261.2025.11249262","is_oa":false,"landing_page_url":"https://doi.org/10.1109/apsipaasc65261.2025.11249262","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","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/A5110156563","display_name":"Chen-Han Wu","orcid":null},"institutions":[{"id":"https://openalex.org/I154864474","display_name":"National Taiwan University of Science and Technology","ror":"https://ror.org/00q09pe49","country_code":"TW","type":"education","lineage":["https://openalex.org/I154864474"]}],"countries":["TW"],"is_corresponding":false,"raw_author_name":"Chen-Han Wu","raw_affiliation_strings":["National Taiwan University of Science and Technology,Taiwan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Taiwan University of Science and Technology,Taiwan","institution_ids":["https://openalex.org/I154864474"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100714445","display_name":"Kuan-Yu Chen","orcid":"https://orcid.org/0000-0001-9656-7551"},"institutions":[{"id":"https://openalex.org/I154864474","display_name":"National Taiwan University of Science and Technology","ror":"https://ror.org/00q09pe49","country_code":"TW","type":"education","lineage":["https://openalex.org/I154864474"]}],"countries":["TW"],"is_corresponding":false,"raw_author_name":"Kuan-Yu Chen","raw_affiliation_strings":["National Taiwan University of Science and Technology,Taiwan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Taiwan University of Science and Technology,Taiwan","institution_ids":["https://openalex.org/I154864474"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I154864474"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.17290304,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"974","last_page":"979"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.6015999913215637,"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":0.6015999913215637,"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":0.10220000147819519,"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/T10028","display_name":"Topic Modeling","score":0.08179999887943268,"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/set","display_name":"Set (abstract data type)","score":0.7111999988555908},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.6360999941825867},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6028000116348267},{"id":"https://openalex.org/keywords/focus","display_name":"Focus (optics)","score":0.4846999943256378},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.4699999988079071}],"concepts":[{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.7111999988555908},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6952000260353088},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.6360999941825867},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6028000116348267},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.492000013589859},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.4846999943256378},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.4699999988079071},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.46549999713897705},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.35019999742507935},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.2994999885559082},{"id":"https://openalex.org/C87007009","wikidata":"https://www.wikidata.org/wiki/Q210832","display_name":"Statistical hypothesis testing","level":2,"score":0.25679999589920044},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.25519999861717224}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/apsipaasc65261.2025.11249262","is_oa":false,"landing_page_url":"https://doi.org/10.1109/apsipaasc65261.2025.11249262","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","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":21,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W2963242190","https://openalex.org/W3024308166","https://openalex.org/W3141961557","https://openalex.org/W3161048756","https://openalex.org/W3196313137","https://openalex.org/W3197797812","https://openalex.org/W4221160683","https://openalex.org/W4226521565","https://openalex.org/W4253028564","https://openalex.org/W4297841327","https://openalex.org/W4312096480","https://openalex.org/W4319862680","https://openalex.org/W4385822602","https://openalex.org/W4402111608","https://openalex.org/W4402112422","https://openalex.org/W4405633161","https://openalex.org/W4406461542","https://openalex.org/W4406461552","https://openalex.org/W4408355411","https://openalex.org/W4408565363"],"related_works":[],"abstract_inverted_index":{"In":[0,104],"high-performing":[1],"automatic":[2],"speech":[3],"recognition":[4,102],"(ASR)":[5],"systems,":[6],"post-processing":[7,27,55,162,209],"plays":[8],"a":[9,117,134,149],"critical":[10],"role":[11],"alongside":[12],"the":[13,32,43,46,78,92,95,110,123,141,189,207],"ASR":[14,83,181,208],"model":[15,38],"itself.":[16],"Recent":[17],"studies":[18,75],"increasingly":[19],"utilize":[20],"n-best":[21],"hypothesis":[22,93,113,155],"lists":[23],"as":[24,165],"inputs":[25],"to":[26,127,147,193,211],"models.":[28],"While":[29],"much":[30],"of":[31,45,80,112,120,145,191],"existing":[33],"research":[34],"focuses":[35],"on":[36,90,109,177],"improving":[37],"architectures,":[39],"we":[40,107,174,187],"observe":[41],"that":[42,139],"composition":[44],"<tex":[47],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[48],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$n$</tex>-best":[49],"list":[50,126],"itself":[51],"can":[52,202],"significantly":[53],"impact":[54],"outcomes.":[56],"Large":[57],"language":[58],"models":[59],"(LLMs)":[60],"have":[61,76,87],"demonstrated":[62],"strong":[63],"performance":[64],"across":[65],"various":[66],"speech-related":[67],"tasks,":[68],"often":[69],"outperforming":[70],"traditional":[71],"methods.":[72],"Although":[73],"some":[74],"explored":[77],"use":[79],"LLMs":[81,146,192,201],"for":[82],"post-processing,":[84],"these":[85],"efforts":[86],"primarily":[88],"focused":[89],"selecting":[91],"with":[94],"lowest":[96],"error":[97,168],"rate":[98],"or":[99],"directly":[100,194],"correcting":[101],"errors.":[103],"this":[105],"paper,":[106],"focus":[108],"task":[111],"set":[114,159],"refinement":[115],"(HSR)-selecting":[116],"high-quality":[118],"subset":[119],"hypotheses":[121],"from":[122],"original":[124],"candidate":[125],"better":[128],"support":[129],"downstream":[130],"processing.":[131],"We":[132],"propose":[133],"novel":[135],"LLM-based":[136],"HSR":[137],"method":[138],"leverages":[140],"semantic":[142],"reasoning":[143],"capabilities":[144],"select":[148],"more":[150],"informative,":[151],"diverse,":[152],"and":[153,167,184,197],"error-tolerant":[154],"set.":[156],"This":[157],"refined":[158],"facilitates":[160],"subsequent":[161],"methods":[163],"such":[164],"reranking":[166,196],"correction.":[169],"To":[170],"validate":[171],"our":[172],"approach,":[173],"conduct":[175],"experiments":[176],"two":[178],"widely":[179],"used":[180],"benchmarks:":[182],"AISHELL-1":[183],"LibriSpeech.":[185],"Furthermore,":[186],"investigate":[188],"potential":[190],"perform":[195],"correction,":[198],"demonstrating":[199],"how":[200],"be":[203],"effectively":[204],"employed":[205],"throughout":[206],"pipeline":[210],"enhance":[212],"overall":[213],"system":[214],"performance.":[215]},"counts_by_year":[],"updated_date":"2026-06-26T08:34:08.712188","created_date":"2025-11-28T00:00:00"}