{"id":"https://openalex.org/W4366197771","doi":"https://doi.org/10.48550/arxiv.2304.06941","title":"AUTOSPARSE: Towards Automated Sparse Training of Deep Neural Networks","display_name":"AUTOSPARSE: Towards Automated Sparse Training of Deep Neural Networks","publication_year":2023,"publication_date":"2023-04-14","ids":{"openalex":"https://openalex.org/W4366197771","doi":"https://doi.org/10.48550/arxiv.2304.06941"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2304.06941","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2304.06941","pdf_url":"https://arxiv.org/pdf/2304.06941","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2304.06941","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5002404356","display_name":"Abhisek Kundu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kundu, Abhisek","raw_affiliation_strings":["Parallel Computing Lab , Intel Labs"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Parallel Computing Lab , Intel Labs","institution_ids":["https://openalex.org/I1343180700"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023420478","display_name":"Naveen Mellempudi","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mellempudi, Naveen K.","raw_affiliation_strings":["Parallel Computing Lab , Intel Labs"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Parallel Computing Lab , Intel Labs","institution_ids":["https://openalex.org/I1343180700"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024653773","display_name":"Dharma Teja Vooturi","orcid":"https://orcid.org/0000-0003-1617-5385","observed_orcids":["https://orcid.org/0000-0003-1617-5385"]},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Vooturi, Dharma Teja","raw_affiliation_strings":["Parallel Computing Lab , Intel Labs"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Parallel Computing Lab , Intel Labs","institution_ids":["https://openalex.org/I1343180700"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5083700279","display_name":"Bharat Kaul","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kaul, Bharat","raw_affiliation_strings":["Parallel Computing Lab , Intel Labs"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Parallel Computing Lab , Intel Labs","institution_ids":["https://openalex.org/I1343180700"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5032238070","display_name":"Pradeep Kumar Dubey","orcid":"https://orcid.org/0000-0001-5853-0619","observed_orcids":["https://orcid.org/0000-0001-5853-0619"]},"institutions":[{"id":"https://openalex.org/I1343180700","display_name":"Intel (United States)","ror":"https://ror.org/01ek73717","country_code":"US","type":"company","lineage":["https://openalex.org/I1343180700"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Dubey, Pradeep","raw_affiliation_strings":["Parallel Computing Lab , Intel Labs"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Parallel Computing Lab , Intel Labs","institution_ids":["https://openalex.org/I1343180700"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I1343180700"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9994999766349792,"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":0.9994999766349792,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.00019999999494757503,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.0,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/sparse-training","display_name":"sparse training","score":0.9290000200271606},{"id":"https://openalex.org/keywords/gradient-masking","display_name":"gradient masking","score":0.5070000290870667},{"id":"https://openalex.org/keywords/flops-reduction","display_name":"FLOPs reduction","score":0.7250000238418579},{"id":"https://openalex.org/keywords/resnet-50","display_name":"ResNet50","score":0.878000020980835},{"id":"https://openalex.org/keywords/mobilenetv1","display_name":"MobileNetV1","score":0.9879999756813049},{"id":"https://openalex.org/keywords/imagenet-1k","display_name":"ImageNet-1K","score":0.8040000200271606},{"id":"https://openalex.org/keywords/mest","display_name":"MEST","score":0.628000020980835}],"concepts":[{"id":"https://openalex.org/C3826847","wikidata":"https://www.wikidata.org/wiki/Q188768","display_name":"FLOPS","level":2,"score":0.9333124160766602},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7601766586303711},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6864737272262573},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.6603831052780151},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.5961238741874695},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.5841405391693115},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5622192025184631},{"id":"https://openalex.org/C126980161","wikidata":"https://www.wikidata.org/wiki/Q863783","display_name":"Simulated annealing","level":2,"score":0.4674715995788574},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.45831745862960815},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4562361240386963},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3473278284072876},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.07394370436668396},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C6557445","wikidata":"https://www.wikidata.org/wiki/Q173113","display_name":"Agronomy","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2304.06941","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2304.06941","pdf_url":"https://arxiv.org/pdf/2304.06941","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2304.06941","doi":"https://doi.org/10.48550/arxiv.2304.06941","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2304.06941","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2304.06941","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2304.06941","pdf_url":"https://arxiv.org/pdf/2304.06941","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4366197771.pdf","grobid_xml":"https://content.openalex.org/works/W4366197771.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2907034307","https://openalex.org/W4414491954","https://openalex.org/W4404649841","https://openalex.org/W7160320482","https://openalex.org/W7167808963","https://openalex.org/W4283729879","https://openalex.org/W4283799372","https://openalex.org/W4207057190","https://openalex.org/W3035180000","https://openalex.org/W3173989853"],"abstract_inverted_index":{"Sparse":[0],"training":[1,14,89,149],"is":[2],"emerging":[3],"as":[4],"a":[5,56],"promising":[6],"avenue":[7],"for":[8,71,106,121,137],"reducing":[9],"the":[10,30,37,69,79],"computational":[11],"cost":[12],"of":[13,33,49],"neural":[15],"networks.":[16],"Several":[17],"recent":[18],"studies":[19],"have":[20],"proposed":[21],"pruning":[22,82,104],"methods":[23,83,105],"using":[24],"learnable":[25,81,103],"thresholds":[26],"to":[27,84],"efficiently":[28],"explore":[29],"non-uniform":[31],"distribution":[32],"sparsity":[34,65],"inherent":[35],"within":[36],"models.":[38],"In":[39],"this":[40],"paper,":[41],"we":[42],"propose":[43],"Gradient":[44],"Annealing":[45],"(GA),":[46],"where":[47,144],"gradients":[48],"masked":[50],"weights":[51],"are":[52],"scaled":[53],"down":[54],"in":[55,118],"non-linear":[57],"manner.":[58],"GA":[59,77],"provides":[60],"an":[61,86],"elegant":[62],"trade-off":[63],"between":[64],"and":[66,109,151],"accuracy":[67,96],"without":[68],"need":[70],"additional":[72],"sparsity-inducing":[73],"regularization.":[74],"We":[75],"integrated":[76],"with":[78,141],"latest":[80],"create":[85],"automated":[87],"sparse":[88,107,139],"algorithm":[90],"called":[91],"AutoSparse,":[92],"which":[93],"achieves":[94,114],"better":[95],"and/or":[97],"training/inference":[98],"FLOPS":[99,120,150],"reduction":[100,117],"than":[101],"existing":[102],"ResNet50":[108,122,140],"MobileNetV1":[110],"on":[111,123],"ImageNet-1K:":[112],"AutoSparse":[113,129],"(2x,":[115],"7x)":[116],"(training,inference)":[119],"ImageNet":[124],"at":[125],"80%":[126,138],"sparsity.":[127],"Finally,":[128],"outperforms":[130],"sparse-to-sparse":[131],"SotA":[132],"method":[133],"MEST":[134,145],"(uniform":[135],"sparsity)":[136],"similar":[142],"accuracy,":[143],"uses":[146],"12%":[147],"more":[148,153],"50%":[152],"inference":[154],"FLOPS.":[155]},"counts_by_year":[{"year":2024,"cited_by_count":1},{"year":2023,"cited_by_count":1}],"updated_date":"2026-10-05T09:50:58.205359","created_date":"2023-04-19T00:00:00"}