{"id":"https://openalex.org/W4408352784","doi":"https://doi.org/10.1109/icassp49660.2025.10890791","title":"Efficient Pruning for Large-Scale Seq2Seq Speech Models without Back-Propagation","display_name":"Efficient Pruning for Large-Scale Seq2Seq Speech Models without Back-Propagation","publication_year":2025,"publication_date":"2025-03-12","ids":{"openalex":"https://openalex.org/W4408352784","doi":"https://doi.org/10.1109/icassp49660.2025.10890791"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49660.2025.10890791","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10890791","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 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/A5114218788","display_name":"Tianteng Gu","orcid":null},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Tianteng Gu","raw_affiliation_strings":["Shanghai Jiao Tong University,Auditory Cognition and Computational Acoustics Lab, MoE Key Lab of Artificial Intelligence,AI Institute Department of Computer Science and Engineering,Shanghai,China"],"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University,Auditory Cognition and Computational Acoustics Lab, MoE Key Lab of Artificial Intelligence,AI Institute Department of Computer Science and Engineering,Shanghai,China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100378232","display_name":"Bei Liu","orcid":"https://orcid.org/0000-0002-5198-0796"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Bei Liu","raw_affiliation_strings":["Shanghai Jiao Tong University,Auditory Cognition and Computational Acoustics Lab, MoE Key Lab of Artificial Intelligence,AI Institute Department of Computer Science and Engineering,Shanghai,China"],"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University,Auditory Cognition and Computational Acoustics Lab, MoE Key Lab of Artificial Intelligence,AI Institute Department of Computer Science and Engineering,Shanghai,China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100341993","display_name":"Yanmin Qian","orcid":"https://orcid.org/0000-0002-0314-3790"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yanmin Qian","raw_affiliation_strings":["Shanghai Jiao Tong University,Auditory Cognition and Computational Acoustics Lab, MoE Key Lab of Artificial Intelligence,AI Institute Department of Computer Science and Engineering,Shanghai,China"],"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University,Auditory Cognition and Computational Acoustics Lab, MoE Key Lab of Artificial Intelligence,AI Institute Department of Computer Science and Engineering,Shanghai,China","institution_ids":["https://openalex.org/I183067930"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5114218788"],"corresponding_institution_ids":["https://openalex.org/I183067930"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.02020961,"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":"5"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9957000017166138,"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":0.9957000017166138,"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":0.9781000018119812,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7017396092414856},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.6159894466400146},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.5916193127632141},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.5336395502090454},{"id":"https://openalex.org/keywords/speech-synthesis","display_name":"Speech synthesis","score":0.4273262619972229},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3778592050075531},{"id":"https://openalex.org/keywords/physics","display_name":"Physics","score":0.08033677935600281}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7017396092414856},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.6159894466400146},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.5916193127632141},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5336395502090454},{"id":"https://openalex.org/C14999030","wikidata":"https://www.wikidata.org/wiki/Q16346","display_name":"Speech synthesis","level":2,"score":0.4273262619972229},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3778592050075531},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.08033677935600281},{"id":"https://openalex.org/C6557445","wikidata":"https://www.wikidata.org/wiki/Q173113","display_name":"Agronomy","level":1,"score":0.0},{"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}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49660.2025.10890791","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10890791","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 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":27,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W3095410713","https://openalex.org/W3203140070","https://openalex.org/W3209059054","https://openalex.org/W3209984917","https://openalex.org/W4297841557","https://openalex.org/W4319862635","https://openalex.org/W4375869065","https://openalex.org/W4385822766","https://openalex.org/W4385823182","https://openalex.org/W4385823417","https://openalex.org/W4392903288","https://openalex.org/W4392904185","https://openalex.org/W4392910583","https://openalex.org/W4400434308","https://openalex.org/W4402115916","https://openalex.org/W6738957339","https://openalex.org/W6771467084","https://openalex.org/W6780218876","https://openalex.org/W6785749098","https://openalex.org/W6788158735","https://openalex.org/W6796551075","https://openalex.org/W6847363464","https://openalex.org/W6848451824","https://openalex.org/W6852423491","https://openalex.org/W6854094408","https://openalex.org/W6870165730"],"related_works":["https://openalex.org/W2373300491","https://openalex.org/W2395294869","https://openalex.org/W2378744544","https://openalex.org/W2594301978","https://openalex.org/W2379704676","https://openalex.org/W1998810860","https://openalex.org/W4206442282","https://openalex.org/W2384505857","https://openalex.org/W2355171581","https://openalex.org/W4229439743"],"abstract_inverted_index":{"Large-scale":[0],"Seq2Seq":[1],"speech":[2,8,56],"models":[3,39,46,57],"like":[4],"Whisper":[5],"excel":[6],"in":[7,100],"recognition":[9],"but":[10],"are":[11],"limited":[12],"by":[13],"their":[14],"high":[15,78],"computational":[16],"demands,":[17],"making":[18],"them":[19],"difficult":[20],"to":[21,54,115],"be":[22,105],"deployed":[23],"on":[24,45,81],"resource-constrained":[25],"devices.":[26],"This":[27,70],"paper":[28],"introduces":[29],"a":[30,60],"novel":[31],"and":[32,42,58,94],"efficient":[33],"pruning":[34,53],"method":[35,88],"for":[36],"compressing":[37],"these":[38],"without":[40],"retraining":[41],"back-propagation,":[43],"focusing":[44],"with":[47,96,107],"encoder-decoder":[48],"architectures.":[49],"We":[50],"adapt":[51],"layer-wise":[52],"large":[55],"introduce":[59],"mixed":[61],"sparsity":[62],"allocation":[63],"strategy":[64],"that":[65],"uses":[66],"only":[67],"forward":[68],"propagation.":[69],"approach":[71],"effectively":[72],"reduces":[73],"model":[74,109,118],"size":[75],"while":[76],"maintaining":[77],"performance.":[79],"Evaluated":[80],"the":[82],"Whisper-large-v3":[83],"across":[84],"various":[85],"datasets,":[86],"our":[87],"can":[89],"almost":[90],"maintain":[91],"Whisper\u2019s":[92],"performance":[93],"robustness":[95],"about":[97],"60%":[98],"reduction":[99],"parameters.":[101],"It":[102],"could":[103],"also":[104],"combined":[106],"other":[108],"compression":[110],"methods":[111],"such":[112],"as":[113],"distillation":[114],"further":[116],"reducing":[117],"size.":[119]},"counts_by_year":[],"updated_date":"2025-12-22T23:10:17.713674","created_date":"2025-10-10T00:00:00"}