{"id":"https://openalex.org/W7165121292","doi":"https://doi.org/10.1145/3787109.3816377","title":"PRISM: Pruning via Rectified-gradient Importance and Saliency Mapping \u2014 making models sparse for execution on edge","display_name":"PRISM: Pruning via Rectified-gradient Importance and Saliency Mapping \u2014 making models sparse for execution on edge","publication_year":2026,"publication_date":"2026-06-18","ids":{"openalex":"https://openalex.org/W7165121292","doi":"https://doi.org/10.1145/3787109.3816377"},"language":null,"primary_location":{"id":"doi:10.1145/3787109.3816377","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3787109.3816377","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Great Lakes Symposium on VLSI 2026","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3787109.3816377","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5034405597","display_name":"Zuxiong Tan","orcid":null},"institutions":[{"id":"https://openalex.org/I84218800","display_name":"University of California, Davis","ror":"https://ror.org/05rrcem69","country_code":"US","type":"education","lineage":["https://openalex.org/I84218800"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zuxiong Tan","raw_affiliation_strings":["University of California, Davis, Davis, CA, USA"],"raw_orcid":"https://orcid.org/0009-0007-9581-8348","affiliations":[{"raw_affiliation_string":"University of California, Davis, Davis, CA, USA","institution_ids":["https://openalex.org/I84218800"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013688249","display_name":"Ali Karkehabadi","orcid":"https://orcid.org/0000-0001-6075-0880"},"institutions":[{"id":"https://openalex.org/I84218800","display_name":"University of California, Davis","ror":"https://ror.org/05rrcem69","country_code":"US","type":"education","lineage":["https://openalex.org/I84218800"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ali Karkehabadi","raw_affiliation_strings":["Electrical and Computer Engineering, University of California, Davis, Davis, CA, USA"],"raw_orcid":"https://orcid.org/0000-0001-6075-0880","affiliations":[{"raw_affiliation_string":"Electrical and Computer Engineering, University of California, Davis, Davis, CA, USA","institution_ids":["https://openalex.org/I84218800"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047382437","display_name":"Houman Homayoun","orcid":"https://orcid.org/0000-0001-8904-4699"},"institutions":[{"id":"https://openalex.org/I84218800","display_name":"University of California, Davis","ror":"https://ror.org/05rrcem69","country_code":"US","type":"education","lineage":["https://openalex.org/I84218800"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Houman Homayoun","raw_affiliation_strings":["University of California, Davis, Davis, CA, USA"],"raw_orcid":"https://orcid.org/0000-0001-8904-4699","affiliations":[{"raw_affiliation_string":"University of California, Davis, Davis, CA, USA","institution_ids":["https://openalex.org/I84218800"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051431139","display_name":"Tooraj Nikoubin","orcid":"https://orcid.org/0000-0003-1724-3503"},"institutions":[{"id":"https://openalex.org/I162577319","display_name":"The University of Texas at Dallas","ror":"https://ror.org/049emcs32","country_code":"US","type":"education","lineage":["https://openalex.org/I162577319"]},{"id":"https://openalex.org/I87573096","display_name":"University of North Texas at Dallas","ror":"https://ror.org/01n51v443","country_code":"US","type":"education","lineage":["https://openalex.org/I87573096"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tooraj Nikoubin","raw_affiliation_strings":["ECE, University of Texas at Dallas, Dallas, TX, USA"],"raw_orcid":"https://orcid.org/0000-0003-1724-3503","affiliations":[{"raw_affiliation_string":"ECE, University of Texas at Dallas, Dallas, TX, USA","institution_ids":["https://openalex.org/I162577319","https://openalex.org/I87573096"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5060036961","display_name":"Avesta Sasan","orcid":"https://orcid.org/0000-0002-4052-8075"},"institutions":[{"id":"https://openalex.org/I84218800","display_name":"University of California, Davis","ror":"https://ror.org/05rrcem69","country_code":"US","type":"education","lineage":["https://openalex.org/I84218800"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Avesta Sasan","raw_affiliation_strings":["University of California, Davis, Davis, CA, USA"],"raw_orcid":"https://orcid.org/0000-0002-4052-8075","affiliations":[{"raw_affiliation_string":"University of California, Davis, Davis, CA, USA","institution_ids":["https://openalex.org/I84218800"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.94328386,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"200","last_page":"207"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.6833999752998352,"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":0.6833999752998352,"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/T11605","display_name":"Visual Attention and Saliency Detection","score":0.12049999833106995,"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/T10273","display_name":"IoT and Edge/Fog Computing","score":0.04100000113248825,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/pruning","display_name":"Pruning","score":0.7601000070571899},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.6139000058174133},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.5011000037193298},{"id":"https://openalex.org/keywords/face","display_name":"Face (sociological concept)","score":0.4864000082015991},{"id":"https://openalex.org/keywords/prism","display_name":"Prism","score":0.45680001378059387},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.3864000141620636},{"id":"https://openalex.org/keywords/edge-device","display_name":"Edge device","score":0.38269999623298645}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7807999849319458},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.7601000070571899},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6794000267982483},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.6139000058174133},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.5011000037193298},{"id":"https://openalex.org/C2779304628","wikidata":"https://www.wikidata.org/wiki/Q3503480","display_name":"Face (sociological concept)","level":2,"score":0.4864000082015991},{"id":"https://openalex.org/C67666897","wikidata":"https://www.wikidata.org/wiki/Q165896","display_name":"Prism","level":2,"score":0.45680001378059387},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.3864000141620636},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.38269999623298645},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.37860000133514404},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.36039999127388},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3368000090122223},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.31310001015663147},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.29910001158714294},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2953000068664551},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.2921000123023987},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.2856000065803528},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2732999920845032},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2517000138759613},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.25029999017715454}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3787109.3816377","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3787109.3816377","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Great Lakes Symposium on VLSI 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3787109.3816377","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3787109.3816377","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Great Lakes Symposium on VLSI 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[{"score":0.4507887065410614,"display_name":"Industry, innovation and infrastructure","id":"https://metadata.un.org/sdg/9"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":29,"referenced_works":["https://openalex.org/W2108598243","https://openalex.org/W2194775991","https://openalex.org/W2736953746","https://openalex.org/W2962851801","https://openalex.org/W2962858109","https://openalex.org/W3011684164","https://openalex.org/W3034251466","https://openalex.org/W3034408420","https://openalex.org/W3034513523","https://openalex.org/W3173390350","https://openalex.org/W3193895134","https://openalex.org/W4213068901","https://openalex.org/W4214633470","https://openalex.org/W4229452442","https://openalex.org/W4283811336","https://openalex.org/W4285505364","https://openalex.org/W4312900176","https://openalex.org/W4313398029","https://openalex.org/W4319083468","https://openalex.org/W4319602321","https://openalex.org/W4380520162","https://openalex.org/W4386071870","https://openalex.org/W4386072014","https://openalex.org/W4386575191","https://openalex.org/W4397046516","https://openalex.org/W4399206618","https://openalex.org/W4402727041","https://openalex.org/W4402917213","https://openalex.org/W4405909471"],"related_works":[],"abstract_inverted_index":{"Deep":[0],"vision":[1],"models":[2],"routinely":[3],"exceed":[4],"the":[5,96],"memory":[6],"and":[7,38,66,81,102,117,152,169,178],"latency":[8],"budgets":[9],"of":[10],"edge":[11],"devices,":[12],"making":[13],"pruning":[14,51],"a":[15,22,49,72],"practical":[16],"necessity.":[17],"However,":[18],"existing":[19],"approaches":[20],"face":[21],"three-way":[23],"trade-off:":[24],"methods":[25],"tailored":[26],"to":[27,62,172],"specific":[28],"architectures":[29],"lack":[30],"generality,":[31],"hardware-friendly":[32],"structured":[33,103,155],"sparsity":[34,104,130],"can":[35,82],"hurt":[36],"accuracy,":[37],"accurate":[39],"importance":[40,70],"estimates":[41],"are":[42],"often":[43],"computationally":[44],"expensive.":[45],"We":[46],"present":[47],"PRISM,":[48],"saliency-based":[50],"framework":[52],"that":[53],"resolves":[54],"this":[55],"tension":[56],"by":[57,115,119,124],"using":[58],"gated":[59],"(rectified)":[60],"gradients":[61],"denoise":[63],"per-sample":[64],"signals":[65],"produce":[67],"reliable":[68],"weight-level":[69],"in":[71],"single":[73],"backward":[74],"pass.":[75],"These":[76],"scores":[77],"accumulate":[78],"over":[79],"data":[80],"be":[83],"aggregated":[84],"along":[85],"structural":[86],"axes\u2014channels,":[87],"neurons,":[88],"attention":[89],"heads,":[90],"or":[91,139],"fixed":[92],"N:":[93],"M":[94],"blocks\u2014so":[95],"same":[97],"criterion":[98],"supports":[99],"both":[100],"unstructured":[101],"with":[105,148,161],"linear-time":[106],"scoring.":[107],"On":[108,135],"ImageNet-1K,":[109],"PRISM":[110,137,164],"prunes":[111],"ResNet-50,":[112],"reducing":[113],"parameters":[114],"41.9%":[116],"MACs":[118],"51.2%":[120],"while":[121],"improving":[122],"Top-1":[123,145],"+0.66":[125],"percentage":[126],"points;":[127],"under":[128],"2:4":[129],"it":[131],"reaches":[132],"78.2%":[133],"Top-1.":[134],"Transformers,":[136],"matches":[138],"surpasses":[140],"strong":[141],"baselines,":[142],"e.g.,":[143],"74.1%":[144],"on":[146],"DeiT-Tiny":[147],"2:":[149],"4":[150],"sparsity,":[151],"outperforms":[153],"prior":[154],"methods.":[156],"By":[157],"coupling":[158],"rectified-gradient":[159],"saliency":[160],"lightweight":[162],"aggregation,":[163],"delivers":[165],"an":[166],"architecture-agnostic,":[167],"hardware-aligned,":[168],"interpretable":[170],"route":[171],"efficient":[173],"deep":[174],"learning":[175],"across":[176],"CNNs":[177],"ViTs.":[179]},"counts_by_year":[],"updated_date":"2026-06-19T15:51:49.773706","created_date":"2026-06-19T00:00:00"}