{"id":"https://openalex.org/W4285247940","doi":"https://doi.org/10.1109/tvlsi.2022.3175582","title":"THETA: A High-Efficiency Training Accelerator for DNNs With Triple-Side Sparsity Exploration","display_name":"THETA: A High-Efficiency Training Accelerator for DNNs With Triple-Side Sparsity Exploration","publication_year":2022,"publication_date":"2022-05-23","ids":{"openalex":"https://openalex.org/W4285247940","doi":"https://doi.org/10.1109/tvlsi.2022.3175582"},"language":"en","primary_location":{"id":"doi:10.1109/tvlsi.2022.3175582","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tvlsi.2022.3175582","pdf_url":null,"source":{"id":"https://openalex.org/S37538908","display_name":"IEEE Transactions on Very Large Scale Integration (VLSI) Systems","issn_l":"1063-8210","issn":["1063-8210","1557-9999"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Very Large Scale Integration (VLSI) Systems","raw_type":"journal-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/A5059107361","display_name":"Jinming Lu","orcid":"https://orcid.org/0000-0002-7134-6514"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Jinming Lu","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005380910","display_name":"Jian Huang","orcid":"https://orcid.org/0000-0002-1125-671X"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jian Huang","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100696999","display_name":"Zhongfeng Wang","orcid":"https://orcid.org/0000-0002-7227-4786"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhongfeng Wang","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5059107361"],"corresponding_institution_ids":["https://openalex.org/I881766915"],"apc_list":null,"apc_paid":null,"fwci":1.3086,"has_fulltext":false,"cited_by_count":13,"citation_normalized_percentile":{"value":0.8123148,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":98},"biblio":{"volume":"30","issue":"8","first_page":"1034","last_page":"1046"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9984999895095825,"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"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9984999895095825,"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"}},{"id":"https://openalex.org/T12676","display_name":"Machine Learning and ELM","score":0.995199978351593,"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/T13918","display_name":"Advanced Data and IoT Technologies","score":0.9775999784469604,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7096344232559204},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.5698720216751099},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.5572041869163513},{"id":"https://openalex.org/keywords/dataflow","display_name":"Dataflow","score":0.543989360332489},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.5297566056251526},{"id":"https://openalex.org/keywords/edge-device","display_name":"Edge device","score":0.49010714888572693},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.47039517760276794},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.4434784948825836},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.41508060693740845},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3890608847141266},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.38240110874176025},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.35221049189567566},{"id":"https://openalex.org/keywords/cloud-computing","display_name":"Cloud computing","score":0.1190791130065918}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7096344232559204},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.5698720216751099},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.5572041869163513},{"id":"https://openalex.org/C96324660","wikidata":"https://www.wikidata.org/wiki/Q205446","display_name":"Dataflow","level":2,"score":0.543989360332489},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.5297566056251526},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.49010714888572693},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.47039517760276794},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.4434784948825836},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.41508060693740845},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3890608847141266},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.38240110874176025},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.35221049189567566},{"id":"https://openalex.org/C79974875","wikidata":"https://www.wikidata.org/wiki/Q483639","display_name":"Cloud computing","level":2,"score":0.1190791130065918},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"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}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tvlsi.2022.3175582","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tvlsi.2022.3175582","pdf_url":null,"source":{"id":"https://openalex.org/S37538908","display_name":"IEEE Transactions on Very Large Scale Integration (VLSI) Systems","issn_l":"1063-8210","issn":["1063-8210","1557-9999"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Very Large Scale Integration (VLSI) Systems","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.7599999904632568,"display_name":"Affordable and clean energy","id":"https://metadata.un.org/sdg/7"}],"awards":[{"id":"https://openalex.org/G1941337620","display_name":null,"funder_award_id":"021014380065","funder_id":"https://openalex.org/F4320335787","funder_display_name":"Fundamental Research Funds for the Central Universities"},{"id":"https://openalex.org/G4112626697","display_name":null,"funder_award_id":"61774082","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320335787","display_name":"Fundamental Research Funds for the Central Universities","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":55,"referenced_works":["https://openalex.org/W2194775991","https://openalex.org/W2289252105","https://openalex.org/W2725159389","https://openalex.org/W2798170643","https://openalex.org/W2888885397","https://openalex.org/W2898985762","https://openalex.org/W2899481200","https://openalex.org/W2910507280","https://openalex.org/W2917518248","https://openalex.org/W2921918777","https://openalex.org/W2945146780","https://openalex.org/W2950656546","https://openalex.org/W2962298324","https://openalex.org/W2964988320","https://openalex.org/W2979439447","https://openalex.org/W2988246957","https://openalex.org/W3003315945","https://openalex.org/W3012561096","https://openalex.org/W3013601031","https://openalex.org/W3015729306","https://openalex.org/W3016542674","https://openalex.org/W3034368386","https://openalex.org/W3035016149","https://openalex.org/W3035332806","https://openalex.org/W3040850704","https://openalex.org/W3047872496","https://openalex.org/W3048842964","https://openalex.org/W3084623993","https://openalex.org/W3088312824","https://openalex.org/W3091592563","https://openalex.org/W3092345941","https://openalex.org/W3092357178","https://openalex.org/W3092581319","https://openalex.org/W3104263540","https://openalex.org/W3105802176","https://openalex.org/W3134304371","https://openalex.org/W3135859967","https://openalex.org/W3137841151","https://openalex.org/W3138154797","https://openalex.org/W3158634533","https://openalex.org/W3164217046","https://openalex.org/W3168545914","https://openalex.org/W3174469560","https://openalex.org/W3185047697","https://openalex.org/W3196171820","https://openalex.org/W4210823275","https://openalex.org/W4240168186","https://openalex.org/W4287118909","https://openalex.org/W6739901393","https://openalex.org/W6751979845","https://openalex.org/W6762580018","https://openalex.org/W6762656248","https://openalex.org/W6772013979","https://openalex.org/W6780482815","https://openalex.org/W6796815506"],"related_works":["https://openalex.org/W3033233036","https://openalex.org/W2973622361","https://openalex.org/W3176282186","https://openalex.org/W4387489555","https://openalex.org/W3185576471","https://openalex.org/W4288024917","https://openalex.org/W4293053895","https://openalex.org/W2983364019","https://openalex.org/W2998183476","https://openalex.org/W3215372595"],"abstract_inverted_index":{"Training":[0],"deep":[1],"neural":[2],"networks":[3],"(DNNs)":[4],"on":[5,25,189],"edge":[6,27],"devices":[7,28],"has":[8],"attracted":[9],"increasing":[10],"attention":[11],"in":[12,39,51,123,208],"real-world":[13],"applications":[14],"for":[15],"domain":[16],"adaption":[17],"and":[18,36,67,109,138,170,187,205,210,224,255],"privacy":[19],"protection.":[20],"However,":[21],"deploying":[22],"DNN":[23],"training":[24,49,99,175,222],"resource-limited":[26],"is":[29,72,78,127,157,177],"challenging":[30],"as":[31,165],"there":[32],"are":[33,64],"massive":[34],"computations":[35],"data":[37,62,107],"transportation":[38],"training.":[40,150],"To":[41,92],"address":[42],"this":[43,52],"issue,":[44],"we":[45,101],"propose":[46],"an":[47,152],"energy-efficient":[48],"accelerator":[50,176],"work":[53],"by":[54],"employing":[55],"a":[56,86,104,110,116,124,131],"hybrid":[57],"compression":[58],"strategy.":[59],"Here,":[60],"various":[61],"redundancies":[63],"fully":[65,139],"exploited,":[66],"the":[68,75,148,174,241],"real":[69],"triple-side":[70,94],"sparsity":[71],"achieved.":[73],"Hence,":[74],"computational":[76,145],"complexity":[77],"drastically":[79],"reduced":[80],"with":[81,143,240],"negligible":[82],"accuracy":[83],"loss":[84],"across":[85],"range":[87],"of":[88],"transfer":[89],"learning":[90],"tasks.":[91],"facilitate":[93],"zero-skipping":[95],"operations":[96,201],"during":[97],"different":[98],"stages,":[100],"first":[102],"present":[103],"novel":[105],"sparse":[106,117],"representation":[108],"triple-sparsity":[111],"index":[112],"matching":[113],"scheme.":[114],"Second,":[115],"tensor":[118],"processing":[119,154],"unit":[120,155,168],"(STPU)":[121],"arranged":[122],"hierarchical":[125],"structure":[126],"developed,":[128],"which":[129],"enables":[130],"flexible":[132],"dataflow":[133],"to":[134,159],"process":[135,186],"convolutional":[136],"(Conv)":[137],"connected":[140],"(FC)":[141],"layers":[142],"diverse":[144],"patterns":[146],"throughout":[147],"entire":[149],"Third,":[151],"auxiliary":[153],"(APU)":[156],"designed":[158],"execute":[160],"some":[161],"postprocessing":[162],"operations,":[163],"such":[164],"rectified":[166],"linear":[167],"(ReLU)":[169],"on-the-fly":[171],"pruning.":[172],"Finally,":[173],"implemented":[178],"under":[179],"Taiwan":[180],"Semiconductor":[181],"Manufacturing":[182],"Company":[183],"(TSMC)":[184],"28-nm":[185],"evaluated":[188],"multiple":[190],"benchmarks.":[191],"The":[192],"experimental":[193],"results":[194],"show":[195],"that":[196],"THETA":[197],"achieves":[198],"7.28\u201322.32":[199],"tera":[200],"per":[202],"second":[203],"(TOPS)":[204],"45.24\u2013133.70":[206],"TOPS/W":[207],"performance":[209],"energy":[211,233,263],"efficiency,":[212,264],"reducing":[213],"40\u2013":[214],"<inline-formula":[215,226,247,256],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[216,227,248,257],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">":[217,228,249,258],"<tex-math":[218,229,250,259],"notation=\"LaTeX\">$72\\times":[219],"$":[220,231,252,261],"</tex-math></inline-formula>":[221,232,253,262],"time":[223],"19\u2013":[225],"notation=\"LaTeX\">$63\\times":[230],"consumption":[234],"over":[235],"dense":[236],"training,":[237],"respectively.":[238,265],"Compared":[239],"prior":[242],"art,":[243],"our":[244],"design":[245],"offers":[246],"notation=\"LaTeX\">$1.6\\times":[251],"throughput":[254],"notation=\"LaTeX\">$1.9\\times":[260]},"counts_by_year":[{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":3},{"year":2023,"cited_by_count":5},{"year":2022,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}