{"id":"https://openalex.org/W3207460419","doi":"https://doi.org/10.48550/arxiv.2110.10811","title":"HALP: Hardware-Aware Latency Pruning","display_name":"HALP: Hardware-Aware Latency Pruning","publication_year":2021,"publication_date":"2021-10-20","ids":{"openalex":"https://openalex.org/W3207460419","doi":"https://doi.org/10.48550/arxiv.2110.10811","mag":"3207460419"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2110.10811","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2110.10811","pdf_url":"https://arxiv.org/pdf/2110.10811","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2110.10811","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5079403120","display_name":"Maying Shen","orcid":"https://orcid.org/0009-0000-9416-680X","observed_orcids":["https://orcid.org/0009-0000-9416-680X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shen, Maying","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5002444694","display_name":"Hongxu Yin","orcid":"https://orcid.org/0000-0002-6481-6389","observed_orcids":["https://orcid.org/0000-0002-6481-6389"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yin, Hongxu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066945976","display_name":"Pavlo A. Molchanov","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Molchanov, Pavlo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103959357","display_name":"Lei Mao","orcid":"https://orcid.org/0009-0002-3004-7195","observed_orcids":["https://orcid.org/0009-0002-3004-7195"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mao, Lei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5062161706","display_name":"Jianna Liu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Jianna","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5101540588","display_name":"Jos\u00e9 Manuel \u00c1lvarez","orcid":"https://orcid.org/0000-0002-7535-6322","observed_orcids":["https://orcid.org/0000-0002-7535-6322"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Alvarez, Jose M.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":15,"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":1.0,"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":1.0,"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":0.0,"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"}},{"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/structural-pruning","display_name":"structural pruning","score":0.7770000100135803},{"id":"https://openalex.org/keywords/optimal-resource-allocation","display_name":"resource allocation optimization","score":0.8149999976158142},{"id":"https://openalex.org/keywords/accuracy-efficiency-trade-off","display_name":"accuracy-efficiency trade-off","score":0.8960000276565552},{"id":"https://openalex.org/keywords/network-throughput","display_name":"network throughput","score":0.5189999938011169},{"id":"https://openalex.org/keywords/ssd-object-detection","display_name":"SSD object detection","score":0.640999972820282},{"id":"https://openalex.org/keywords/image-classification","display_name":"image classification","score":0.5370000004768372}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7385751605033875},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.7265120148658752},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.6083089113235474},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4259381890296936},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.4147590398788452},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3796141743659973},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2673640847206116},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.18664517998695374},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.07561904191970825},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","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":2,"locations":[{"id":"pmh:oai:arXiv.org:2110.10811","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2110.10811","pdf_url":"https://arxiv.org/pdf/2110.10811","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2110.10811","doi":"https://doi.org/10.48550/arxiv.2110.10811","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2110.10811","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2110.10811","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2110.10811","pdf_url":"https://arxiv.org/pdf/2110.10811","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3207460419.pdf","grobid_xml":"https://content.openalex.org/works/W3207460419.grobid-xml"},"referenced_works_count":52,"referenced_works":["https://openalex.org/W639708223","https://openalex.org/W1667652561","https://openalex.org/W1686810756","https://openalex.org/W1821462560","https://openalex.org/W2031489346","https://openalex.org/W2117539524","https://openalex.org/W2119144962","https://openalex.org/W2495425901","https://openalex.org/W2515385951","https://openalex.org/W2554931888","https://openalex.org/W2557728737","https://openalex.org/W2560017826","https://openalex.org/W2612445135","https://openalex.org/W2707890836","https://openalex.org/W2808938483","https://openalex.org/W2883780447","https://openalex.org/W2886851211","https://openalex.org/W2886953980","https://openalex.org/W2894740066","https://openalex.org/W2895260923","https://openalex.org/W2896409484","https://openalex.org/W2899771611","https://openalex.org/W2902342446","https://openalex.org/W2915589364","https://openalex.org/W2939504873","https://openalex.org/W2946948417","https://openalex.org/W2947681860","https://openalex.org/W2955419283","https://openalex.org/W2962861284","https://openalex.org/W2963351448","https://openalex.org/W2963813662","https://openalex.org/W2964233199","https://openalex.org/W2967733054","https://openalex.org/W2970632972","https://openalex.org/W2979429416","https://openalex.org/W2980137827","https://openalex.org/W2984618279","https://openalex.org/W2995492258","https://openalex.org/W2995895045","https://openalex.org/W3007623470","https://openalex.org/W3011514561","https://openalex.org/W3016265891","https://openalex.org/W3030163527","https://openalex.org/W3034290785","https://openalex.org/W3034819536","https://openalex.org/W3034940165","https://openalex.org/W3034957837","https://openalex.org/W3035332806","https://openalex.org/W3092358983","https://openalex.org/W3106250896","https://openalex.org/W3109212549","https://openalex.org/W3111652984"],"related_works":["https://openalex.org/W7133204290","https://openalex.org/W4378509443","https://openalex.org/W4392458085","https://openalex.org/W3160783538","https://openalex.org/W4405180165","https://openalex.org/W4229452442","https://openalex.org/W7169661559","https://openalex.org/W4401069990","https://openalex.org/W4399836838","https://openalex.org/W4388843595"],"abstract_inverted_index":{"Structural":[0],"pruning":[1,19,76,102,128,145],"can":[2,63],"simplify":[3],"network":[4,133],"architecture":[5],"and":[6,53,104,113,121],"improve":[7],"inference":[8],"speed.":[9],"We":[10,107],"propose":[11],"Hardware-Aware":[12],"Latency":[13],"Pruning":[14],"(HALP)":[15],"that":[16],"formulates":[17],"structural":[18,75],"as":[20],"a":[21,36,78,155],"global":[22,54,74],"resource":[23],"allocation":[24],"optimization":[25],"problem,":[26],"aiming":[27],"at":[28],"maximizing":[29],"the":[30,87],"accuracy":[31,59,140],"while":[32],"constraining":[33],"latency":[34,45,50],"under":[35,77],"predefined":[37],"budget.":[38],"For":[39,143],"filter":[40],"importance":[41],"ranking,":[42],"HALP":[43,96,109,131,148,159],"leverages":[44],"lookup":[46],"table":[47],"to":[48,57,72,97],"track":[49],"reduction":[51],"potential":[52],"saliency":[55],"score":[56],"gauge":[58],"drop.":[60,158],"Both":[61],"metrics":[62],"be":[64],"evaluated":[65],"very":[66],"efficiently":[67],"during":[68],"pruning,":[69],"allowing":[70],"us":[71],"reformulate":[73],"reward":[79],"maximization":[80],"problem":[81,88],"given":[82],"target":[83],"constraint.":[84],"This":[85],"makes":[86],"solvable":[89],"via":[90],"our":[91],"augmented":[92],"knapsack":[93],"solver,":[94],"enabling":[95],"surpass":[98],"prior":[99,162],"work":[100],"in":[101],"efficacy":[103],"accuracy-efficiency":[105],"trade-off.":[106],"examine":[108],"on":[110,119,129,146],"both":[111],"classification":[112],"detection":[114],"tasks,":[115],"over":[116],"varying":[117],"networks,":[118],"ImageNet":[120],"VOC":[122],"datasets.":[123],"In":[124],"particular,":[125],"for":[126],"ResNet-50/-101":[127],"ImageNet,":[130],"improves":[132,149],"throughput":[134,150],"by":[135,151,165],"$1.60\\times$/$1.90\\times$":[136],"with":[137,153],"$+0.3\\%$/$-0.2\\%$":[138],"top-1":[139],"changes,":[141],"respectively.":[142],"SSD":[144],"VOC,":[147],"$1.94\\times$":[152],"only":[154],"$0.56$":[156],"mAP":[157],"consistently":[160],"outperforms":[161],"art,":[163],"sometimes":[164],"large":[166],"margins.":[167]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":1},{"year":2024,"cited_by_count":3},{"year":2023,"cited_by_count":6},{"year":2022,"cited_by_count":3}],"updated_date":"2026-11-10T11:39:43.021295","created_date":"2021-10-25T00:00:00"}