{"id":"https://openalex.org/W4384026302","doi":"https://doi.org/10.1109/ets56758.2023.10173972","title":"Understanding Permanent Hardware Failures in Deep Learning Training Accelerator Systems","display_name":"Understanding Permanent Hardware Failures in Deep Learning Training Accelerator Systems","publication_year":2023,"publication_date":"2023-05-22","ids":{"openalex":"https://openalex.org/W4384026302","doi":"https://doi.org/10.1109/ets56758.2023.10173972"},"language":"en","primary_location":{"id":"doi:10.1109/ets56758.2023.10173972","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ets56758.2023.10173972","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2023 IEEE European Test Symposium (ETS)","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/A5100702362","display_name":"Yi He","orcid":"https://orcid.org/0000-0001-7206-4845"},"institutions":[{"id":"https://openalex.org/I40347166","display_name":"University of Chicago","ror":"https://ror.org/024mw5h28","country_code":"US","type":"education","lineage":["https://openalex.org/I40347166"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Yi He","raw_affiliation_strings":["University of Chicago,Chicago,USA","University of Chicago, Chicago, USA"],"affiliations":[{"raw_affiliation_string":"University of Chicago,Chicago,USA","institution_ids":["https://openalex.org/I40347166"]},{"raw_affiliation_string":"University of Chicago, Chicago, USA","institution_ids":["https://openalex.org/I40347166"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101989053","display_name":"Yanjing Li","orcid":"https://orcid.org/0000-0003-0124-0463"},"institutions":[{"id":"https://openalex.org/I40347166","display_name":"University of Chicago","ror":"https://ror.org/024mw5h28","country_code":"US","type":"education","lineage":["https://openalex.org/I40347166"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yanjing Li","raw_affiliation_strings":["University of Chicago,Chicago,USA","University of Chicago, Chicago, USA"],"affiliations":[{"raw_affiliation_string":"University of Chicago,Chicago,USA","institution_ids":["https://openalex.org/I40347166"]},{"raw_affiliation_string":"University of Chicago, Chicago, USA","institution_ids":["https://openalex.org/I40347166"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5100702362"],"corresponding_institution_ids":["https://openalex.org/I40347166"],"apc_list":null,"apc_paid":null,"fwci":0.6174,"has_fulltext":false,"cited_by_count":5,"citation_normalized_percentile":{"value":0.69055793,"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":"1","last_page":"6"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9979000091552734,"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":0.9979000091552734,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9973000288009644,"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/T11005","display_name":"Radiation Effects in Electronics","score":0.9972000122070312,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7647926807403564},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.603091835975647},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5711972117424011},{"id":"https://openalex.org/keywords/fault-injection","display_name":"Fault injection","score":0.5675773620605469},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.5371170043945312},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5264504551887512},{"id":"https://openalex.org/keywords/resilience","display_name":"Resilience (materials science)","score":0.4741998314857483},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.46106964349746704},{"id":"https://openalex.org/keywords/software","display_name":"Software","score":0.40950027108192444},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.37342965602874756},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.17750588059425354}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7647926807403564},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.603091835975647},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5711972117424011},{"id":"https://openalex.org/C2775928411","wikidata":"https://www.wikidata.org/wiki/Q2041312","display_name":"Fault injection","level":3,"score":0.5675773620605469},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5371170043945312},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5264504551887512},{"id":"https://openalex.org/C2779585090","wikidata":"https://www.wikidata.org/wiki/Q3457762","display_name":"Resilience (materials science)","level":2,"score":0.4741998314857483},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.46106964349746704},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.40950027108192444},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.37342965602874756},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.17750588059425354},{"id":"https://openalex.org/C97355855","wikidata":"https://www.wikidata.org/wiki/Q11473","display_name":"Thermodynamics","level":1,"score":0.0},{"id":"https://openalex.org/C151730666","wikidata":"https://www.wikidata.org/wiki/Q7205","display_name":"Paleontology","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"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":1,"locations":[{"id":"doi:10.1109/ets56758.2023.10173972","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ets56758.2023.10173972","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2023 IEEE European Test Symposium (ETS)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Peace, Justice and strong institutions","score":0.8199999928474426,"id":"https://metadata.un.org/sdg/16"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":31,"referenced_works":["https://openalex.org/W1677182931","https://openalex.org/W1972649107","https://openalex.org/W2081714388","https://openalex.org/W2164529645","https://openalex.org/W2186094539","https://openalex.org/W2194775991","https://openalex.org/W2257408573","https://openalex.org/W2964108906","https://openalex.org/W2978384356","https://openalex.org/W2996011786","https://openalex.org/W3012897490","https://openalex.org/W3014034225","https://openalex.org/W3105084904","https://openalex.org/W3118608800","https://openalex.org/W3119922080","https://openalex.org/W3120931476","https://openalex.org/W3171842021","https://openalex.org/W3189398641","https://openalex.org/W3190062760","https://openalex.org/W3205165858","https://openalex.org/W3214851476","https://openalex.org/W4226335648","https://openalex.org/W4293584584","https://openalex.org/W4385245566","https://openalex.org/W4386212580","https://openalex.org/W6631943919","https://openalex.org/W6686583229","https://openalex.org/W6750227808","https://openalex.org/W6776283548","https://openalex.org/W6787972765","https://openalex.org/W6810666586"],"related_works":["https://openalex.org/W2767601850","https://openalex.org/W3014207222","https://openalex.org/W2617887951","https://openalex.org/W2246057640","https://openalex.org/W2120051590","https://openalex.org/W2538450276","https://openalex.org/W1894599085","https://openalex.org/W2112914176","https://openalex.org/W2029915748","https://openalex.org/W2416955034"],"abstract_inverted_index":{"Hardware":[0],"failures":[1,46,54,69,127,155],"pose":[2],"critical":[3],"threats":[4],"to":[5,102,150,159],"deep":[6,90],"neural":[7,191],"network":[8,192],"(DNN)":[9],"training":[10,74,93,129,135],"workloads,":[11],"and":[12,110,131,147,174,183],"the":[13,21,35,45,65,81,105,119],"urgency":[14],"of":[15,44,64,107,180],"tackling":[16],"this":[17,77,86,139],"challenge":[18,25],"(known":[19],"as":[20],"Silent":[22],"Data":[23],"Corruption":[24],"in":[26,48,55],"a":[27,41,60,98],"broader":[28],"context)":[29],"has":[30],"been":[31],"raised":[32],"widely":[33],"by":[34],"industry.":[36],"Based":[37,137],"on":[38,72,85,89,122,138,164],"industry":[39],"reports,":[40],"large":[42],"number":[43],"observed":[47],"real":[49],"systems":[50],"are":[51,157,172],"permanent":[52,108,125,153],"hardware":[53,126,154],"logic.":[56],"However,":[57],"there":[58],"is":[59],"very":[61],"limited":[62],"understanding":[63,121],"effects":[66,106],"that":[67,156,169],"these":[68],"can":[70],"impose":[71],"DNN":[73],"workloads.":[75],"In":[76],"paper,":[78],"we":[79,142],"present":[80],"first":[82],"resilience":[83],"study":[84],"subject,":[87],"focusing":[88],"learning":[91],"(DL)":[92],"accelerator":[94],"systems.":[95],"We":[96],"developed":[97,143],"fault":[99,113],"injection":[100,114],"framework":[101],"accurately":[103],"simulate":[104],"faults,":[109],"conducted":[111],"100K":[112],"experiments.":[115],"Our":[116],"results":[117],"provide":[118],"fundamental":[120],"how":[123],"logic":[124,152],"affect":[128],"workloads":[130],"eventually":[132],"generate":[133,160],"unexpected":[134,161],"outcomes.":[136,162],"new":[140],"knowledge,":[141],"efficient":[144],"software-based":[145],"detection":[146],"recovery":[148],"techniques":[149,171],"mitigate":[151],"likely":[158],"Evaluation":[163],"Google":[165],"Cloud":[166],"TPUs":[167],"shows":[168],"our":[170],"effective":[173],"practical:":[175],"they":[176],"require":[177],"15\u221225":[178],"lines":[179],"code":[181],"change,":[182],"introduce":[184],"0.004%\u22120.025%":[185],"performance/energy":[186],"overhead":[187],"for":[188],"various":[189],"representative":[190],"models.":[193]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}