{"id":"https://openalex.org/W3007929985","doi":"https://doi.org/10.1145/3373087.3375385","title":"V-LSTM: An Efficient LSTM Accelerator Using Fixed Nonzero-Ratio Viterbi-Based Pruning","display_name":"V-LSTM: An Efficient LSTM Accelerator Using Fixed Nonzero-Ratio Viterbi-Based Pruning","publication_year":2020,"publication_date":"2020-02-23","ids":{"openalex":"https://openalex.org/W3007929985","doi":"https://doi.org/10.1145/3373087.3375385","mag":"3007929985"},"language":"en","primary_location":{"id":"doi:10.1145/3373087.3375385","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3373087.3375385","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2020 ACM/SIGDA International Symposium on Field-Programmable Gate Arrays","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/A5101792302","display_name":"Taesu Kim","orcid":"https://orcid.org/0000-0001-7987-527X"},"institutions":[{"id":"https://openalex.org/I123900574","display_name":"Pohang University of Science and Technology","ror":"https://ror.org/04xysgw12","country_code":"KR","type":"education","lineage":["https://openalex.org/I123900574"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Taesu Kim","raw_affiliation_strings":["Pohang University of Science and Technology, Pohang, South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Pohang University of Science and Technology, Pohang, South Korea","institution_ids":["https://openalex.org/I123900574"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023936391","display_name":"Daehyun Ahn","orcid":"https://orcid.org/0000-0001-8050-9963"},"institutions":[{"id":"https://openalex.org/I123900574","display_name":"Pohang University of Science and Technology","ror":"https://ror.org/04xysgw12","country_code":"KR","type":"education","lineage":["https://openalex.org/I123900574"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Daehyun Ahn","raw_affiliation_strings":["Pohang University of Science and Technology, Pohang, South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Pohang University of Science and Technology, Pohang, South Korea","institution_ids":["https://openalex.org/I123900574"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5003219699","display_name":"Jae\u2010Joon Kim","orcid":"https://orcid.org/0000-0001-5175-8258"},"institutions":[{"id":"https://openalex.org/I123900574","display_name":"Pohang University of Science and Technology","ror":"https://ror.org/04xysgw12","country_code":"KR","type":"education","lineage":["https://openalex.org/I123900574"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jae-Joon Kim","raw_affiliation_strings":["Pohang University of Science and Technology, Pohang, South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Pohang University of Science and Technology, Pohang, South Korea","institution_ids":["https://openalex.org/I123900574"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.4062,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":{"value":0.68270434,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"326","last_page":"326"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9955999851226807,"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":0.9955999851226807,"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/T12326","display_name":"Network Packet Processing and Optimization","score":0.9914000034332275,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9908999800682068,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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":0.8429299592971802},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.526874840259552},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.5207579135894775},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4360731840133667},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.4168986678123474},{"id":"https://openalex.org/keywords/memory-footprint","display_name":"Memory footprint","score":0.4159744083881378},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.4158228933811188},{"id":"https://openalex.org/keywords/hardware-acceleration","display_name":"Hardware acceleration","score":0.4138634502887726},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4122288227081299},{"id":"https://openalex.org/keywords/viterbi-algorithm","display_name":"Viterbi algorithm","score":0.41219621896743774},{"id":"https://openalex.org/keywords/recurrent-neural-network","display_name":"Recurrent neural network","score":0.41181179881095886},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.3654389977455139},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.3618156313896179},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.32160690426826477},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.2954925298690796},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.28252550959587097},{"id":"https://openalex.org/keywords/hidden-markov-model","display_name":"Hidden Markov model","score":0.22059929370880127}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8429299592971802},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.526874840259552},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.5207579135894775},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4360731840133667},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.4168986678123474},{"id":"https://openalex.org/C74912251","wikidata":"https://www.wikidata.org/wiki/Q6815727","display_name":"Memory footprint","level":2,"score":0.4159744083881378},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4158228933811188},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.4138634502887726},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4122288227081299},{"id":"https://openalex.org/C60582962","wikidata":"https://www.wikidata.org/wiki/Q83886","display_name":"Viterbi algorithm","level":3,"score":0.41219621896743774},{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.41181179881095886},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.3654389977455139},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.3618156313896179},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.32160690426826477},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.2954925298690796},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.28252550959587097},{"id":"https://openalex.org/C23224414","wikidata":"https://www.wikidata.org/wiki/Q176769","display_name":"Hidden Markov model","level":2,"score":0.22059929370880127},{"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/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3373087.3375385","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3373087.3375385","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2020 ACM/SIGDA International Symposium on Field-Programmable Gate Arrays","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4298287631","https://openalex.org/W2953061907","https://openalex.org/W1847088711","https://openalex.org/W4225394202","https://openalex.org/W3036642985","https://openalex.org/W3032952384","https://openalex.org/W2160451571","https://openalex.org/W2495256954","https://openalex.org/W2259317772","https://openalex.org/W4283700541"],"abstract_inverted_index":{"Long":[0],"Short-Term":[1],"Memory":[2],"(LSTM)":[3],"has":[4,71],"been":[5,72],"widely":[6],"adopted":[7],"in":[8,45,120],"tasks":[9],"with":[10,51,106],"sequence":[11],"data,":[12],"such":[13],"as":[14],"speech":[15],"recognition":[16],"and":[17,39,56,65,80,131],"language":[18],"modeling.":[19],"LSTM":[20,47,68,102,147],"brought":[21],"significant":[22],"accuracy":[23,108],"improvement":[24],"by":[25,104],"introducing":[26],"additional":[27,112],"parameters":[28,38],"to":[29,43,61,167],"Recurrent":[30],"Neural":[31],"Network":[32],"(RNN).":[33],"However,":[34],"increasing":[35],"number":[36],"of":[37,67,101,117,158,175],"computations":[40],"also":[41],"led":[42],"inefficiency":[44],"computing":[46],"on":[48,114,150,183],"edge":[49],"devices":[50],"limited":[52],"on-chip":[53],"memory":[54,99],"size":[55],"DRAM":[57],"bandwidth.":[58],"In":[59,84],"order":[60],"reduce":[62,97],"the":[63,90,98,115,123,128,134,142,151,159,164],"latency":[64,172],"energy":[66],"computations,":[69],"there":[70],"a":[73],"pressing":[74],"need":[75],"for":[76,192],"model":[77],"compression":[78,156],"schemes":[79],"suitable":[81],"hardware":[82],"accelerators.":[83,177],"this":[85],"paper,":[86],"we":[87,140],"first":[88],"propose":[89,141],"Fixed":[91],"Nonzero-ratio":[92],"Viterbi-based":[93,121],"Pruning,":[94,122],"which":[95],"can":[96],"footprint":[100],"models":[103],"96%":[105],"negligible":[107],"loss.":[109],"By":[110],"applying":[111],"constraints":[113],"distribution":[116],"surviving":[118],"weights":[119],"proposed":[124,152,160,165,179],"pruning":[125,153,161],"scheme":[126,162],"mitigates":[127],"load-imbalance":[129],"problem":[130],"thereby":[132],"increases":[133],"processing":[135],"engine":[136],"utilization":[137],"rate.":[138],"Then,":[139],"V-LSTM,":[143],"an":[144],"efficient":[145],"sparse":[146],"accelerator":[148,166,180],"based":[149],"scheme.":[154],"High":[155],"ratio":[157],"allows":[163],"achieve":[168],"24.9%":[169],"lower":[170],"per-sample":[171],"than":[173],"that":[174],"state-of-the-art":[176],"The":[178],"is":[181],"implemented":[182],"Xilinx":[184],"VC-709":[185],"FPGA":[186],"evaluation":[187],"board":[188],"running":[189],"at":[190],"200MHz":[191],"evaluation.":[193]},"counts_by_year":[{"year":2023,"cited_by_count":1},{"year":2022,"cited_by_count":2}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}