{"id":"https://openalex.org/W4224929675","doi":"https://doi.org/10.1109/icassp43922.2022.9746885","title":"Mixed Precision DNN Quantization for Overlapped Speech Separation and Recognition","display_name":"Mixed Precision DNN Quantization for Overlapped Speech Separation and Recognition","publication_year":2022,"publication_date":"2022-04-27","ids":{"openalex":"https://openalex.org/W4224929675","doi":"https://doi.org/10.1109/icassp43922.2022.9746885"},"language":"en","primary_location":{"id":"doi:10.1109/icassp43922.2022.9746885","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp43922.2022.9746885","pdf_url":null,"source":{"id":"https://openalex.org/S4363607702","display_name":"ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2022 - 2022 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/A5011790590","display_name":"Junhao Xu","orcid":"https://orcid.org/0000-0002-9650-0290"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":true,"raw_author_name":"Junhao Xu","raw_affiliation_strings":["The Chinese University of Hong Kong"],"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004643540","display_name":"Jianwei Yu","orcid":"https://orcid.org/0000-0002-2449-1436"},"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":"Jianwei Yu","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/A5037109470","display_name":"Xunying Liu","orcid":"https://orcid.org/0000-0001-6725-1160"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Xunying Liu","raw_affiliation_strings":["The Chinese University of Hong Kong"],"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5019458385","display_name":"Helen Meng","orcid":"https://orcid.org/0000-0002-4427-3532"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Helen Meng","raw_affiliation_strings":["The Chinese University of Hong Kong"],"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong","institution_ids":["https://openalex.org/I177725633"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5011790590"],"corresponding_institution_ids":["https://openalex.org/I177725633"],"apc_list":null,"apc_paid":null,"fwci":1.2158,"has_fulltext":false,"cited_by_count":10,"citation_normalized_percentile":{"value":0.80824937,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"7297","last_page":"7301"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10860","display_name":"Speech and Audio Processing","score":1.0,"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"}},"topics":[{"id":"https://openalex.org/T10860","display_name":"Speech and Audio Processing","score":1.0,"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"}},{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","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/T11309","display_name":"Music and Audio Processing","score":0.9987000226974487,"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/quantization","display_name":"Quantization (signal processing)","score":0.7790862321853638},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7193824648857117},{"id":"https://openalex.org/keywords/pesq","display_name":"PESQ","score":0.6247342824935913},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.5596184134483337},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.4787789583206177},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.42821285128593445},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.3978321850299835},{"id":"https://openalex.org/keywords/speech-enhancement","display_name":"Speech enhancement","score":0.37809351086616516},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3122932016849518},{"id":"https://openalex.org/keywords/noise-reduction","display_name":"Noise reduction","score":0.1078898012638092}],"concepts":[{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.7790862321853638},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7193824648857117},{"id":"https://openalex.org/C103734657","wikidata":"https://www.wikidata.org/wiki/Q2739975","display_name":"PESQ","level":4,"score":0.6247342824935913},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5596184134483337},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4787789583206177},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.42821285128593445},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3978321850299835},{"id":"https://openalex.org/C2776182073","wikidata":"https://www.wikidata.org/wiki/Q7575395","display_name":"Speech enhancement","level":3,"score":0.37809351086616516},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3122932016849518},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.1078898012638092}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp43922.2022.9746885","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp43922.2022.9746885","pdf_url":null,"source":{"id":"https://openalex.org/S4363607702","display_name":"ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.5899999737739563,"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":61,"referenced_works":["https://openalex.org/W1677182931","https://openalex.org/W1902934009","https://openalex.org/W2057450058","https://openalex.org/W2060108923","https://openalex.org/W2102764819","https://openalex.org/W2148613904","https://openalex.org/W2168729028","https://openalex.org/W2291877678","https://openalex.org/W2300242332","https://openalex.org/W2405920868","https://openalex.org/W2531409750","https://openalex.org/W2739789140","https://openalex.org/W2763421725","https://openalex.org/W2803431233","https://openalex.org/W2885311373","https://openalex.org/W2889503488","https://openalex.org/W2890952074","https://openalex.org/W2891205112","https://openalex.org/W2901739041","https://openalex.org/W2909607850","https://openalex.org/W2939360348","https://openalex.org/W2939497224","https://openalex.org/W2944972166","https://openalex.org/W2952218014","https://openalex.org/W2952265507","https://openalex.org/W2972693890","https://openalex.org/W2982041622","https://openalex.org/W3001133453","https://openalex.org/W3008011454","https://openalex.org/W3011424113","https://openalex.org/W3015199127","https://openalex.org/W3016230677","https://openalex.org/W3034309359","https://openalex.org/W3048889251","https://openalex.org/W3096408984","https://openalex.org/W3097522836","https://openalex.org/W3097798370","https://openalex.org/W3099330747","https://openalex.org/W3160348851","https://openalex.org/W3161201807","https://openalex.org/W3163368926","https://openalex.org/W3164605550","https://openalex.org/W3196364802","https://openalex.org/W3217016993","https://openalex.org/W4288026258","https://openalex.org/W6639703010","https://openalex.org/W6684472137","https://openalex.org/W6696676114","https://openalex.org/W6698200048","https://openalex.org/W6714058667","https://openalex.org/W6745245109","https://openalex.org/W6753278433","https://openalex.org/W6754392867","https://openalex.org/W6754420807","https://openalex.org/W6756104738","https://openalex.org/W6761289755","https://openalex.org/W6762182681","https://openalex.org/W6770425567","https://openalex.org/W6771595542","https://openalex.org/W6776096719","https://openalex.org/W6803829408"],"related_works":["https://openalex.org/W3016109656","https://openalex.org/W2058482658","https://openalex.org/W1973895194","https://openalex.org/W3135613579","https://openalex.org/W3209446892","https://openalex.org/W2345688450","https://openalex.org/W4386746628","https://openalex.org/W1980687383","https://openalex.org/W2105126898","https://openalex.org/W2792294082"],"abstract_inverted_index":{"Recognition":[0],"of":[1,88,176,181],"overlapped":[2,156],"speech":[3,14,94,157,173],"has":[4],"been":[5],"a":[6,34,89],"highly":[7],"challenging":[8],"task":[9],"to":[10,37,54,65,84,195],"date.":[11],"State-of-the-art":[12],"multi-channel":[13,93],"separation":[15,95,136,174],"system":[16],"are":[17,47,77,102],"becoming":[18],"increasingly":[19],"complex":[20],"and":[21,52,134,183],"expensive":[22],"for":[23,56],"practical":[24],"applications.":[25],"To":[26],"this":[27,69],"end,":[28],"low-bit":[29],"neural":[30,146],"network":[31],"quantization":[32,45,66,75,113,165],"provides":[33],"powerful":[35],"solution":[36],"dramatically":[38],"reduce":[39],"their":[40],"model":[41,63],"size.":[42],"However,":[43],"current":[44],"methods":[46,76],"based":[48,92,116,142],"on":[49,117,143,151],"uniform":[50,170],"precision":[51,73,100,133,145,164,171],"fail":[53],"account":[55],"the":[57,119,128,152,161,169],"varying":[58],"performance":[59],"sensitivity":[60,114],"at":[61],"different":[62],"components":[64,87],"errors.":[67],"In":[68],"paper,":[70],"novel":[71],"mixed":[72,144,163],"DNN":[74],"proposed":[78,162],"by":[79],"applying":[80],"locally":[81],"variable":[82],"bit-widths":[83,178],"individual":[85],"TCN":[86],"TF":[90],"masking":[91],"system.":[96],"The":[97,108,138],"optimal":[98],"local":[99],"settings":[101],"automatically":[103],"learned":[104],"using":[105],"three":[106],"techniques.":[107],"first":[109],"two":[110],"approaches":[111],"utilize":[112],"metrics":[115],"either":[118],"mean":[120],"square":[121],"error":[122,190],"(MSE)":[123],"loss":[124],"function":[125],"curvature,":[126],"or":[127],"KL-divergence":[129],"measured":[130],"between":[131],"full":[132],"quantized":[135],"models.":[137],"third":[139],"approach":[140],"is":[141],"architecture":[147],"search.":[148],"Experiments":[149],"conducted":[150],"LRS3-TED":[153],"corpus":[154],"simulated":[155],"data":[158],"suggest":[159],"that":[160],"techniques":[166],"consistently":[167],"outperform":[168],"baseline":[172],"systems":[175],"comparable":[177],"in":[179],"terms":[180],"SI-SNR":[182],"PESQ":[184],"scores":[185],"as":[186,188],"well":[187],"word":[189],"rate":[191],"(WER)":[192],"reductions":[193],"up":[194],"2.88%":[196],"absolute":[197],"(8%":[198],"relative).":[199]},"counts_by_year":[{"year":2025,"cited_by_count":3},{"year":2024,"cited_by_count":6},{"year":2023,"cited_by_count":1}],"updated_date":"2026-04-04T06:10:10.580331","created_date":"2025-10-10T00:00:00"}