{"id":"https://openalex.org/W4394931310","doi":"https://doi.org/10.1145/3656643","title":"PQA: Exploring the Potential of Product Quantization in DNN Hardware Acceleration","display_name":"PQA: Exploring the Potential of Product Quantization in DNN Hardware Acceleration","publication_year":2024,"publication_date":"2024-04-18","ids":{"openalex":"https://openalex.org/W4394931310","doi":"https://doi.org/10.1145/3656643"},"language":"en","primary_location":{"id":"doi:10.1145/3656643","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3656643","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3656643","source":{"id":"https://openalex.org/S112809824","display_name":"ACM Transactions on Reconfigurable Technology and Systems","issn_l":"1936-7406","issn":["1936-7406","1936-7414"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Reconfigurable Technology and Systems","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"bronze","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3656643","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5069242854","display_name":"Ahmed F. AbouElhamayed","orcid":"https://orcid.org/0000-0001-6381-2936"},"institutions":[{"id":"https://openalex.org/I205783295","display_name":"Cornell University","ror":"https://ror.org/05bnh6r87","country_code":"US","type":"education","lineage":["https://openalex.org/I205783295"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ahmed Abouelhamayed","raw_affiliation_strings":["Electrical and Computer Engineering, Cornell University, New York, United States","Electrical and Computer Engineering, Cornell University, New York, USA"],"raw_orcid":"https://orcid.org/0000-0001-6381-2936","affiliations":[{"raw_affiliation_string":"Electrical and Computer Engineering, Cornell University, New York, United States","institution_ids":["https://openalex.org/I205783295"]},{"raw_affiliation_string":"Electrical and Computer Engineering, Cornell University, New York, USA","institution_ids":["https://openalex.org/I205783295"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111279939","display_name":"Angela Cui","orcid":null},"institutions":[{"id":"https://openalex.org/I205783295","display_name":"Cornell University","ror":"https://ror.org/05bnh6r87","country_code":"US","type":"education","lineage":["https://openalex.org/I205783295"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Angela Cui","raw_affiliation_strings":["Cornell University, Ithaca, United States","Electrical and Computer Engineering, Cornell University, New York, USA"],"raw_orcid":"https://orcid.org/0000-0001-9997-1701","affiliations":[{"raw_affiliation_string":"Cornell University, Ithaca, United States","institution_ids":["https://openalex.org/I205783295"]},{"raw_affiliation_string":"Electrical and Computer Engineering, Cornell University, New York, USA","institution_ids":["https://openalex.org/I205783295"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031116318","display_name":"Javier Fern\u00e1ndez-Marqu\u00e9s","orcid":null},"institutions":[{"id":"https://openalex.org/I2802518283","display_name":"Flower Hospital","ror":"https://ror.org/01xar8k65","country_code":"US","type":"healthcare","lineage":["https://openalex.org/I2802518283"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Javier Fernandez-marques","raw_affiliation_strings":["Flower Labs, Cambridge, United Kingdom of Great Britain and Northern Ireland","Flower Labs, Cambridge, UK"],"raw_orcid":"https://orcid.org/0000-0003-3747-6523","affiliations":[{"raw_affiliation_string":"Flower Labs, Cambridge, United Kingdom of Great Britain and Northern Ireland","institution_ids":["https://openalex.org/I2802518283"]},{"raw_affiliation_string":"Flower Labs, Cambridge, UK","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5045638679","display_name":"Nicholas D. Lane","orcid":"https://orcid.org/0000-0002-2728-8273"},"institutions":[{"id":"https://openalex.org/I241749","display_name":"University of Cambridge","ror":"https://ror.org/013meh722","country_code":"GB","type":"education","lineage":["https://openalex.org/I241749"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Nicholas Lane","raw_affiliation_strings":["Department of Computer Science and Technology, University of Cambridge, Cambridge, United Kingdom of Great Britain and Northern Ireland","Department of Computer Science and Technology, University of Cambridge, Cambridge, UK"],"raw_orcid":"https://orcid.org/0000-0002-2728-8273","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Technology, University of Cambridge, Cambridge, United Kingdom of Great Britain and Northern Ireland","institution_ids":["https://openalex.org/I241749"]},{"raw_affiliation_string":"Department of Computer Science and Technology, University of Cambridge, Cambridge, UK","institution_ids":["https://openalex.org/I241749"]}]},{"author_position":"last","author":{"id":null,"display_name":"Mohamed Abdelfattah","orcid":"https://orcid.org/0000-0002-4568-8932"},"institutions":[{"id":"https://openalex.org/I205783295","display_name":"Cornell University","ror":"https://ror.org/05bnh6r87","country_code":"US","type":"education","lineage":["https://openalex.org/I205783295"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mohamed Abdelfattah","raw_affiliation_strings":["ECE, Cornell University, New York, United States","Cornell University, New York, USA"],"raw_orcid":"https://orcid.org/0000-0002-4568-8932","affiliations":[{"raw_affiliation_string":"ECE, Cornell University, New York, United States","institution_ids":["https://openalex.org/I205783295"]},{"raw_affiliation_string":"Cornell University, New York, USA","institution_ids":["https://openalex.org/I205783295"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.8175,"has_fulltext":true,"cited_by_count":4,"citation_normalized_percentile":{"value":0.71349484,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":97},"biblio":{"volume":"18","issue":"1","first_page":"1","last_page":"29"},"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9995999932289124,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.9991999864578247,"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/computer-science","display_name":"Computer science","score":0.8102096915245056},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.7267903089523315},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.6476863622665405},{"id":"https://openalex.org/keywords/deep-neural-networks","display_name":"Deep neural networks","score":0.5330475568771362},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4901517331600189},{"id":"https://openalex.org/keywords/hardware-acceleration","display_name":"Hardware acceleration","score":0.4803798794746399},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.4378904104232788},{"id":"https://openalex.org/keywords/product","display_name":"Product (mathematics)","score":0.435799777507782},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.4084376096725464},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.4069851040840149},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.3988664150238037},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.39770811796188354},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.28445371985435486},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.1642076075077057},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.0679449737071991}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8102096915245056},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.7267903089523315},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.6476863622665405},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.5330475568771362},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4901517331600189},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.4803798794746399},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4378904104232788},{"id":"https://openalex.org/C90673727","wikidata":"https://www.wikidata.org/wiki/Q901718","display_name":"Product (mathematics)","level":2,"score":0.435799777507782},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4084376096725464},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4069851040840149},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.3988664150238037},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.39770811796188354},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.28445371985435486},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.1642076075077057},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0679449737071991},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3656643","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3656643","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3656643","source":{"id":"https://openalex.org/S112809824","display_name":"ACM Transactions on Reconfigurable Technology and Systems","issn_l":"1936-7406","issn":["1936-7406","1936-7414"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Reconfigurable Technology and Systems","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1145/3656643","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3656643","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3656643","source":{"id":"https://openalex.org/S112809824","display_name":"ACM Transactions on Reconfigurable Technology and Systems","issn_l":"1936-7406","issn":["1936-7406","1936-7414"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Reconfigurable Technology and Systems","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320307102","display_name":"Intel Corporation","ror":"https://ror.org/01ek73717"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4394931310.pdf","grobid_xml":"https://content.openalex.org/works/W4394931310.grobid-xml"},"referenced_works_count":37,"referenced_works":["https://openalex.org/W2054951389","https://openalex.org/W2077815765","https://openalex.org/W2124509324","https://openalex.org/W2148154194","https://openalex.org/W2172654076","https://openalex.org/W2194775991","https://openalex.org/W2300242332","https://openalex.org/W2346804249","https://openalex.org/W2524985544","https://openalex.org/W2547875792","https://openalex.org/W2574797063","https://openalex.org/W2734358244","https://openalex.org/W2797583228","https://openalex.org/W2884288768","https://openalex.org/W2889119508","https://openalex.org/W2892054964","https://openalex.org/W2942998388","https://openalex.org/W2945146780","https://openalex.org/W2953212265","https://openalex.org/W2963122961","https://openalex.org/W2963163009","https://openalex.org/W2964121744","https://openalex.org/W2982479999","https://openalex.org/W2982512683","https://openalex.org/W2989331028","https://openalex.org/W3007581096","https://openalex.org/W3034190694","https://openalex.org/W3034344052","https://openalex.org/W3042416028","https://openalex.org/W3094136258","https://openalex.org/W3198035615","https://openalex.org/W4281991412","https://openalex.org/W4285483879","https://openalex.org/W4293820342","https://openalex.org/W4308083526","https://openalex.org/W4311638103","https://openalex.org/W4313170858"],"related_works":["https://openalex.org/W4205547277","https://openalex.org/W3204400881","https://openalex.org/W3214410901","https://openalex.org/W3204296682","https://openalex.org/W3183118997","https://openalex.org/W2917767146","https://openalex.org/W4319952061","https://openalex.org/W4280636456","https://openalex.org/W4285818394","https://openalex.org/W4388913998"],"abstract_inverted_index":{"Conventional":[0],"multiply-accumulate":[1],"(MAC)":[2],"operations":[3,140],"have":[4],"long":[5],"dominated":[6],"computation":[7],"time":[8],"for":[9,88,169],"deep":[10],"neural":[11,16],"networks":[12,17],"(DNNs),":[13],"especially":[14],"convolutional":[15],"(CNNs).":[18],"Recently,":[19],"product":[20],"quantization":[21],"(PQ)":[22],"has":[23],"been":[24],"applied":[25],"to":[26,34,54,69,92,97,114,141,162],"these":[27],"workloads,":[28],"replacing":[29],"MACs":[30],"with":[31,104,128],"memory":[32],"lookups":[33],"pre-computed":[35],"dot":[36],"products.":[37],"To":[38],"better":[39],"understand":[40],"the":[41,71,136,143,167],"efficiency":[42,148],"tradeoffs":[43,73],"of":[44,74,126,138],"product-quantized":[45],"DNNs":[46],"(PQ-DNNs),":[47],"we":[48,64,118,134,159],"create":[49],"a":[50,98,129],"custom":[51],"hardware":[52,147],"accelerator":[53],"parallelize":[55],"and":[56,60,78,149],"accelerate":[57],"nearest-neighbor":[58],"search":[59],"dot-product":[61],"lookups.":[62],"Additionally,":[63],"perform":[65],"an":[66],"empirical":[67],"study":[68],"investigate":[70,142],"efficiency\u2013accuracy":[72],"different":[75],"PQ":[76,83,116,139],"parameterizations":[77],"training":[79],"methods.":[80],"We":[81],"identify":[82],"configurations":[84],"that":[85],"improve":[86],"performance-per-area":[87,127],"ResNet20":[89],"by":[90,122],"up":[91],"3.1\u00d7,":[93],"even":[94],"when":[95],"compared":[96],"highly":[99],"optimized":[100],"conventional":[101],"DNN":[102,164],"accelerator,":[103],"similar":[105],"improvements":[106],"on":[107,145,155],"two":[108],"additional":[109],"compact":[110,157],"DNNs.":[111],"When":[112],"comparing":[113],"recent":[115],"solutions,":[117],"outperform":[119],"prior":[120],"work":[121],"4\u00d7":[123],"in":[124],"terms":[125],"0.6%":[130],"accuracy":[131,165],"degradation.":[132],"Finally,":[133],"reduce":[135],"bitwidth":[137],"impact":[144],"both":[146],"accuracy.":[150],"With":[151],"only":[152],"2\u20136-bit":[153],"precision":[154],"three":[156],"DNNs,":[158],"were":[160],"able":[161],"maintain":[163],"eliminating":[166],"need":[168],"DSPs.":[170]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":1}],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-10T00:00:00"}