{"id":"https://openalex.org/W3128097214","doi":"https://doi.org/10.1145/3394885.3431533","title":"Efficient Accuracy Recovery in Approximate Neural Networks by Systematic Error Modelling","display_name":"Efficient Accuracy Recovery in Approximate Neural Networks by Systematic Error Modelling","publication_year":2021,"publication_date":"2021-01-18","ids":{"openalex":"https://openalex.org/W3128097214","doi":"https://doi.org/10.1145/3394885.3431533","mag":"3128097214"},"language":"en","primary_location":{"id":"doi:10.1145/3394885.3431533","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3394885.3431533","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 26th Asia and South Pacific Design Automation Conference","raw_type":"proceedings-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/A5069309172","display_name":"Cecilia De la Parra","orcid":"https://orcid.org/0000-0002-1463-6822"},"institutions":[{"id":"https://openalex.org/I889804353","display_name":"Robert Bosch (Germany)","ror":"https://ror.org/01fe0jt45","country_code":"DE","type":"company","lineage":["https://openalex.org/I889804353"]}],"countries":["DE"],"is_corresponding":true,"raw_author_name":"Cecilia De la Parra","raw_affiliation_strings":["Robert Bosch GmbH, Renningen, Germany"],"affiliations":[{"raw_affiliation_string":"Robert Bosch GmbH, Renningen, Germany","institution_ids":["https://openalex.org/I889804353"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5056912495","display_name":"Andre Guntoro","orcid":"https://orcid.org/0000-0003-4144-0283"},"institutions":[{"id":"https://openalex.org/I889804353","display_name":"Robert Bosch (Germany)","ror":"https://ror.org/01fe0jt45","country_code":"DE","type":"company","lineage":["https://openalex.org/I889804353"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Andre Guntoro","raw_affiliation_strings":["Robert Bosch GmbH, Renningen, Germany"],"affiliations":[{"raw_affiliation_string":"Robert Bosch GmbH, Renningen, Germany","institution_ids":["https://openalex.org/I889804353"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100755285","display_name":"Akash Kumar","orcid":"https://orcid.org/0000-0001-7125-1737"},"institutions":[{"id":"https://openalex.org/I78650965","display_name":"TU Dresden","ror":"https://ror.org/042aqky30","country_code":"DE","type":"education","lineage":["https://openalex.org/I78650965"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Akash Kumar","raw_affiliation_strings":["Technische Universit\u00e4t Dresden, Dresden, Germany"],"affiliations":[{"raw_affiliation_string":"Technische Universit\u00e4t Dresden, Dresden, Germany","institution_ids":["https://openalex.org/I78650965"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5069309172"],"corresponding_institution_ids":["https://openalex.org/I889804353"],"apc_list":null,"apc_paid":null,"fwci":0.5014,"has_fulltext":false,"cited_by_count":7,"citation_normalized_percentile":{"value":0.62410737,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"365","last_page":"371"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9998999834060669,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9998999834060669,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.9993000030517578,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9991000294685364,"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":0.7788617610931396},{"id":"https://openalex.org/keywords/retraining","display_name":"Retraining","score":0.7345014810562134},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.6776490211486816},{"id":"https://openalex.org/keywords/deep-neural-networks","display_name":"Deep neural networks","score":0.6257236003875732},{"id":"https://openalex.org/keywords/approximation-error","display_name":"Approximation error","score":0.6162964701652527},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.48440009355545044},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.4515503942966461},{"id":"https://openalex.org/keywords/energy","display_name":"Energy (signal processing)","score":0.42721307277679443},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.38126999139785767},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.3249151110649109},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.1324056088924408},{"id":"https://openalex.org/keywords/wireless","display_name":"Wireless","score":0.07773357629776001},{"id":"https://openalex.org/keywords/statistics","display_name":"Statistics","score":0.05817118287086487}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7788617610931396},{"id":"https://openalex.org/C2778712577","wikidata":"https://www.wikidata.org/wiki/Q3505966","display_name":"Retraining","level":2,"score":0.7345014810562134},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.6776490211486816},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.6257236003875732},{"id":"https://openalex.org/C122383733","wikidata":"https://www.wikidata.org/wiki/Q865920","display_name":"Approximation error","level":2,"score":0.6162964701652527},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.48440009355545044},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.4515503942966461},{"id":"https://openalex.org/C186370098","wikidata":"https://www.wikidata.org/wiki/Q442787","display_name":"Energy (signal processing)","level":2,"score":0.42721307277679443},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.38126999139785767},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3249151110649109},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.1324056088924408},{"id":"https://openalex.org/C555944384","wikidata":"https://www.wikidata.org/wiki/Q249","display_name":"Wireless","level":2,"score":0.07773357629776001},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.05817118287086487},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.0},{"id":"https://openalex.org/C155202549","wikidata":"https://www.wikidata.org/wiki/Q178803","display_name":"International trade","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3394885.3431533","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3394885.3431533","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 26th Asia and South Pacific Design Automation Conference","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Affordable and clean energy","score":0.9100000262260437,"id":"https://metadata.un.org/sdg/7"}],"awards":[{"id":"https://openalex.org/G7276696126","display_name":null,"funder_award_id":"380524764","funder_id":"https://openalex.org/F4320320879","funder_display_name":"Deutsche Forschungsgemeinschaft"}],"funders":[{"id":"https://openalex.org/F4320320879","display_name":"Deutsche Forschungsgemeinschaft","ror":"https://ror.org/018mejw64"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":21,"referenced_works":["https://openalex.org/W1841592590","https://openalex.org/W1974078116","https://openalex.org/W1998917233","https://openalex.org/W2028499920","https://openalex.org/W2111406701","https://openalex.org/W2123045220","https://openalex.org/W2533121491","https://openalex.org/W2612139336","https://openalex.org/W2737121650","https://openalex.org/W2748818695","https://openalex.org/W2798993323","https://openalex.org/W2804268686","https://openalex.org/W2808908725","https://openalex.org/W2871705258","https://openalex.org/W2897792138","https://openalex.org/W2949650786","https://openalex.org/W2958306322","https://openalex.org/W2964299589","https://openalex.org/W2997288154","https://openalex.org/W3036274717","https://openalex.org/W3118608800"],"related_works":["https://openalex.org/W2081982437","https://openalex.org/W4394857231","https://openalex.org/W2027050655","https://openalex.org/W3028244590","https://openalex.org/W4254349500","https://openalex.org/W2014369232","https://openalex.org/W3122042562","https://openalex.org/W2050078012","https://openalex.org/W2060761133","https://openalex.org/W2360307734"],"abstract_inverted_index":{"Approximate":[0],"Computing":[1],"is":[2],"a":[3,44],"promising":[4],"paradigm":[5],"for":[6,47,89,127,133,138,140,158],"mitigating":[7],"the":[8,49,71,80,99,112,121],"computational":[9],"demands":[10],"of":[11,74,123,154],"Deep":[12],"Neural":[13],"Networks":[14],"(DNNs),":[15],"by":[16,29,53,83,102],"leveraging":[17],"DNN":[18,26,100,108,142],"performance":[19],"and":[20,61,78,94,135],"area,":[21],"throughput":[22],"or":[23],"power.":[24],"The":[25],"accuracy,":[27],"affected":[28],"such":[30,84],"approximations,":[31],"can":[32],"be":[33],"then":[34],"effectively":[35],"improved":[36],"through":[37],"retraining.":[38],"In":[39],"this":[40,67],"paper,":[41],"we":[42,69,97],"present":[43],"novel":[45],"methodology":[46],"modelling":[48],"approximation":[50],"error":[51,81,105],"introduced":[52],"approximate":[54,76,149],"hardware":[55],"in":[56],"DNNs,":[57],"which":[58,144],"accelerates":[59],"retraining":[60,129],"achieves":[62],"negligible":[63],"accuracy":[64,113,164],"loss.":[65],"To":[66],"end,":[68],"implement":[70],"behavioral":[72],"simulation":[73],"several":[75],"multipliers":[77,150],"model":[79,106],"generated":[82],"approximations":[85],"on":[86,92],"pre-trained":[87],"DNNs":[88,161],"image":[90],"classification":[91],"CIFAR10":[93,134],"ImageNet.":[95],"Finally,":[96],"optimize":[98],"parameters":[101],"applying":[103],"our":[104,124],"during":[107],"retraining,":[109],"to":[110,116,147,156],"recover":[111],"lost":[114],"due":[115],"approximations.":[117],"Experimental":[118],"results":[119],"demonstrate":[120],"efficiency":[122],"proposed":[125],"method":[126],"accelerated":[128],"(11":[130],"x":[131],"faster":[132,137],"8x":[136],"ImageNet)":[139],"full":[141],"approximation,":[143],"allows":[145],"us":[146],"deploy":[148],"with":[151,162],"energy":[152],"savings":[153],"up":[155],"36%":[157],"8-bit":[159],"precision":[160],"an":[163],"loss":[165],"lower":[166],"than":[167],"1%.":[168]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":1},{"year":2022,"cited_by_count":2}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}