{"id":"https://openalex.org/W2991232789","doi":"https://doi.org/10.1109/tc.2019.2954495","title":"Grow and Prune Compact, Fast, and Accurate LSTMs","display_name":"Grow and Prune Compact, Fast, and Accurate LSTMs","publication_year":2019,"publication_date":"2019-11-20","ids":{"openalex":"https://openalex.org/W2991232789","doi":"https://doi.org/10.1109/tc.2019.2954495","mag":"2991232789"},"language":"en","primary_location":{"id":"doi:10.1109/tc.2019.2954495","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tc.2019.2954495","pdf_url":null,"source":{"id":"https://openalex.org/S157670870","display_name":"IEEE Transactions on Computers","issn_l":"0018-9340","issn":["0018-9340","0016-9340","1557-9956","2326-3814"],"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 Computers","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/A5080137972","display_name":"Xiaoliang Dai","orcid":"https://orcid.org/0000-0003-3098-2714"},"institutions":[{"id":"https://openalex.org/I20089843","display_name":"Princeton University","ror":"https://ror.org/00hx57361","country_code":"US","type":"education","lineage":["https://openalex.org/I20089843"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Xiaoliang Dai","raw_affiliation_strings":["Princeton University, Princeton, USA"],"raw_orcid":"https://orcid.org/0000-0003-3098-2714","affiliations":[{"raw_affiliation_string":"Princeton University, Princeton, USA","institution_ids":["https://openalex.org/I20089843"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5002444694","display_name":"Hongxu Yin","orcid":"https://orcid.org/0000-0002-6481-6389"},"institutions":[{"id":"https://openalex.org/I20089843","display_name":"Princeton University","ror":"https://ror.org/00hx57361","country_code":"US","type":"education","lineage":["https://openalex.org/I20089843"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hongxu Yin","raw_affiliation_strings":["Princeton University, Princeton, USA"],"raw_orcid":"https://orcid.org/0000-0002-6481-6389","affiliations":[{"raw_affiliation_string":"Princeton University, Princeton, USA","institution_ids":["https://openalex.org/I20089843"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5086131079","display_name":"Niraj K. Jha","orcid":"https://orcid.org/0000-0002-1539-0369"},"institutions":[{"id":"https://openalex.org/I20089843","display_name":"Princeton University","ror":"https://ror.org/00hx57361","country_code":"US","type":"education","lineage":["https://openalex.org/I20089843"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Niraj K. Jha","raw_affiliation_strings":["Princeton University, Princeton, USA"],"raw_orcid":"https://orcid.org/0000-0002-1539-0369","affiliations":[{"raw_affiliation_string":"Princeton University, Princeton, USA","institution_ids":["https://openalex.org/I20089843"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I20089843"],"apc_list":null,"apc_paid":null,"fwci":7.6242,"has_fulltext":false,"cited_by_count":91,"citation_normalized_percentile":{"value":0.97759856,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":100},"biblio":{"volume":"69","issue":"3","first_page":"441","last_page":"452"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9987000226974487,"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/T10028","display_name":"Topic Modeling","score":0.9987000226974487,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9983999729156494,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9977999925613403,"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":0.8810592889785767},{"id":"https://openalex.org/keywords/flops","display_name":"FLOPS","score":0.7074030637741089},{"id":"https://openalex.org/keywords/overfitting","display_name":"Overfitting","score":0.6640402674674988},{"id":"https://openalex.org/keywords/closed-captioning","display_name":"Closed captioning","score":0.6531679630279541},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.5779367089271545},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.5199181437492371},{"id":"https://openalex.org/keywords/redundancy","display_name":"Redundancy (engineering)","score":0.5127271413803101},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.49881649017333984},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.4761000871658325},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.46959319710731506},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4508693516254425},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.44321537017822266},{"id":"https://openalex.org/keywords/recurrent-neural-network","display_name":"Recurrent neural network","score":0.4406728148460388},{"id":"https://openalex.org/keywords/word","display_name":"Word (group theory)","score":0.4278007745742798},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.42156949639320374},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.3923832178115845},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.37270474433898926},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.37200891971588135},{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.244637131690979},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.20923036336898804}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8810592889785767},{"id":"https://openalex.org/C3826847","wikidata":"https://www.wikidata.org/wiki/Q188768","display_name":"FLOPS","level":2,"score":0.7074030637741089},{"id":"https://openalex.org/C22019652","wikidata":"https://www.wikidata.org/wiki/Q331309","display_name":"Overfitting","level":3,"score":0.6640402674674988},{"id":"https://openalex.org/C157657479","wikidata":"https://www.wikidata.org/wiki/Q2367247","display_name":"Closed captioning","level":3,"score":0.6531679630279541},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.5779367089271545},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.5199181437492371},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.5127271413803101},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.49881649017333984},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.4761000871658325},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.46959319710731506},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4508693516254425},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.44321537017822266},{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.4406728148460388},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.4278007745742798},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.42156949639320374},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.3923832178115845},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.37270474433898926},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.37200891971588135},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.244637131690979},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.20923036336898804},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","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},{"id":"https://openalex.org/C104317684","wikidata":"https://www.wikidata.org/wiki/Q7187","display_name":"Gene","level":2,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0},{"id":"https://openalex.org/C55493867","wikidata":"https://www.wikidata.org/wiki/Q7094","display_name":"Biochemistry","level":1,"score":0.0},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tc.2019.2954495","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tc.2019.2954495","pdf_url":null,"source":{"id":"https://openalex.org/S157670870","display_name":"IEEE Transactions on Computers","issn_l":"0018-9340","issn":["0018-9340","0016-9340","1557-9956","2326-3814"],"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 Computers","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G2351308602","display_name":null,"funder_award_id":"CNS-1617640","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":87,"referenced_works":["https://openalex.org/W587794757","https://openalex.org/W1522301498","https://openalex.org/W1576227399","https://openalex.org/W1591801644","https://openalex.org/W1686810756","https://openalex.org/W1724438581","https://openalex.org/W1861492603","https://openalex.org/W1895577753","https://openalex.org/W1899504021","https://openalex.org/W1902237438","https://openalex.org/W1905882502","https://openalex.org/W1921523184","https://openalex.org/W1922655562","https://openalex.org/W1924770834","https://openalex.org/W1956340063","https://openalex.org/W1996901117","https://openalex.org/W2005708641","https://openalex.org/W2064675550","https://openalex.org/W2076063813","https://openalex.org/W2101105183","https://openalex.org/W2112984492","https://openalex.org/W2119615570","https://openalex.org/W2127141656","https://openalex.org/W2130942839","https://openalex.org/W2143612262","https://openalex.org/W2157331557","https://openalex.org/W2157590573","https://openalex.org/W2161591461","https://openalex.org/W2163605009","https://openalex.org/W2167215970","https://openalex.org/W2172166488","https://openalex.org/W2193413348","https://openalex.org/W2194775991","https://openalex.org/W2409027918","https://openalex.org/W2509836288","https://openalex.org/W2525778437","https://openalex.org/W2585720638","https://openalex.org/W2769644379","https://openalex.org/W2774000609","https://openalex.org/W2794664194","https://openalex.org/W2798170643","https://openalex.org/W2803021247","https://openalex.org/W2885533908","https://openalex.org/W2899771611","https://openalex.org/W2919115771","https://openalex.org/W2950248853","https://openalex.org/W2952899695","https://openalex.org/W2963212250","https://openalex.org/W2963319203","https://openalex.org/W2963674932","https://openalex.org/W2963803379","https://openalex.org/W2963847166","https://openalex.org/W2963917928","https://openalex.org/W2963991999","https://openalex.org/W2964084166","https://openalex.org/W2964121744","https://openalex.org/W2964301388","https://openalex.org/W2964335273","https://openalex.org/W2980137827","https://openalex.org/W3104263540","https://openalex.org/W4293714597","https://openalex.org/W4297689207","https://openalex.org/W6617368339","https://openalex.org/W6631190155","https://openalex.org/W6631636882","https://openalex.org/W6635446068","https://openalex.org/W6637373629","https://openalex.org/W6637709462","https://openalex.org/W6639102338","https://openalex.org/W6639497327","https://openalex.org/W6640090968","https://openalex.org/W6640185926","https://openalex.org/W6640212811","https://openalex.org/W6679436768","https://openalex.org/W6683826617","https://openalex.org/W6684191040","https://openalex.org/W6684563725","https://openalex.org/W6685405536","https://openalex.org/W6687566353","https://openalex.org/W6714142977","https://openalex.org/W6725588366","https://openalex.org/W6727690538","https://openalex.org/W6736780897","https://openalex.org/W6746200960","https://openalex.org/W6746839373","https://openalex.org/W6756040250","https://openalex.org/W6898505805"],"related_works":["https://openalex.org/W2594897229","https://openalex.org/W2151348424","https://openalex.org/W4221142855","https://openalex.org/W2050138804","https://openalex.org/W4290708361","https://openalex.org/W2129812225","https://openalex.org/W26527944","https://openalex.org/W2523799048","https://openalex.org/W2155620340","https://openalex.org/W2944691285"],"abstract_inverted_index":{"Long":[0],"short-term":[1],"memory":[2],"(LSTM)":[3],"has":[4],"been":[5],"widely":[6],"used":[7],"for":[8,115],"sequential":[9],"data":[10],"modeling.":[11],"Researchers":[12],"have":[13,112],"increased":[14],"LSTM":[15,19,47],"depth":[16],"by":[17,141,146,150,157,179,184,222,227],"stacking":[18],"cells":[20],"to":[21,37,53,84,198,239,247],"improve":[22,153],"performance.":[23],"This":[24,98],"incurs":[25],"model":[26,171,189,214,232],"redundancy,":[27],"increases":[28,61,233],"run-time":[29,76,148,182,225],"delay,":[30],"and":[31,75,93,103,120,152,181,224,251],"makes":[32],"the":[33,71,87,101,104,126,130,137,154,162,166,169,175,191,202,207,212,218,234],"LSTMs":[34],"more":[35],"prone":[36],"overfitting.":[38],"To":[39],"address":[40],"these":[41],"problems,":[42],"we":[43,172,215],"propose":[44],"a":[45],"hidden-layer":[46],"(H-LSTM)":[48],"that":[49],"adds":[50],"hidden":[51,88],"layers":[52,89],"LSTM's":[54],"original":[55],"one-level":[56],"nonlinear":[57],"control":[58,109],"gates.":[59,110],"H-LSTM":[60,108],"accuracy":[62],"while":[63],"employing":[64],"fewer":[65],"external":[66],"stacked":[67],"layers,":[68],"thus":[69],"reducing":[70],"number":[72,138,176,219],"of":[73,96,107,139,177,220],"parameters":[74,140,178,221],"latency":[77,149,183,226],"significantly.":[78],"We":[79,111],"employ":[80],"grow-and-prune":[81],"(GP)":[82],"training":[83],"iteratively":[85],"adjust":[86],"through":[90],"gradient-based":[91],"growth":[92],"magnitude-based":[94],"pruning":[95],"connections.":[97],"learns":[99],"both":[100],"weights":[102],"compact":[105],"architecture":[106,128,164],"GP-trained":[113,242],"H-LSTMs":[114,243],"image":[116],"captioning,":[117],"speech":[118],"recognition,":[119],"neural":[121],"machine":[122],"translation":[123],"applications.":[124],"For":[125,161,201],"NeuralTalk":[127],"on":[129,165,206],"MSCOCO":[131],"dataset,":[132,168,211],"our":[133],"three":[134],"models":[135],"reduce":[136],"38.7\u00d7":[142],"[floating-point":[143],"operations":[144],"(FLOPs)":[145],"45.5\u00d7],":[147],"4.5\u00d7,":[151],"CIDEr-D":[155],"score":[156,236],"2.8":[158],"percent,":[159],"respectively.":[160],"DeepSpeech2":[163],"AN4":[167],"first":[170,213],"generated":[173,216],"reduces":[174,190,217],"19.4\u00d7":[180],"37.4":[185],"percent.":[186,200,229],"The":[187,230],"second":[188,231],"word":[192],"error":[193],"rate":[194],"(WER)":[195],"from":[196,237],"12.9":[197],"8.7":[199],"encoder-decoder":[203],"sequence-to-sequence":[204],"network":[205],"IWSLT":[208],"2014":[209],"German-English":[210],"10.8\u00d7":[223],"14.2":[228],"BLEU":[235],"30.02":[238],"30.98.":[240],"Thus,":[241],"can":[244],"be":[245,248],"seen":[246],"compact,":[249],"fast,":[250],"accurate.":[252]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":10},{"year":2024,"cited_by_count":15},{"year":2023,"cited_by_count":9},{"year":2022,"cited_by_count":21},{"year":2021,"cited_by_count":16},{"year":2020,"cited_by_count":12},{"year":2019,"cited_by_count":5}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}