{"id":"https://openalex.org/W7133897824","doi":"https://doi.org/10.48550/arxiv.2603.03491","title":"When Small Variations Become Big Failures: Reliability Challenges in Compute-in-Memory Neural Accelerators","display_name":"When Small Variations Become Big Failures: Reliability Challenges in Compute-in-Memory Neural Accelerators","publication_year":2026,"publication_date":"2026-03-03","ids":{"openalex":"https://openalex.org/W7133897824","doi":"https://doi.org/10.48550/arxiv.2603.03491"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2603.03491","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"type":"article","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Qin, Yifan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qin, Yifan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128215775","display_name":"Jiahao Zheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Jiahao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128196738","display_name":"Zheyu Yan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yan, Zheyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112538942","display_name":"Wujie Wen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wen, Wujie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128129359","display_name":"Xiaobo Sharon Hu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Xiaobo Sharon","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5123197509","display_name":"Yiyu Shi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shi, Yiyu","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":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.19463946,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.510200023651123,"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":0.510200023651123,"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":0.1606999933719635,"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":0.07940000295639038,"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/reliability","display_name":"Reliability (semiconductor)","score":0.6647999882698059},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.6021999716758728},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.5881999731063843},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5726000070571899},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.4275999963283539},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.42640000581741333},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.421999990940094},{"id":"https://openalex.org/keywords/efficient-energy-use","display_name":"Efficient energy use","score":0.4097999930381775}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7148000001907349},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.6647999882698059},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.6021999716758728},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.5881999731063843},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5726000070571899},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4677000045776367},{"id":"https://openalex.org/C200601418","wikidata":"https://www.wikidata.org/wiki/Q2193887","display_name":"Reliability engineering","level":1,"score":0.4410000145435333},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4311000108718872},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.4275999963283539},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.42640000581741333},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.421999990940094},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.4097999930381775},{"id":"https://openalex.org/C80469333","wikidata":"https://www.wikidata.org/wiki/Q189088","display_name":"Von Neumann architecture","level":2,"score":0.34790000319480896},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.3384000062942505},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.325300008058548},{"id":"https://openalex.org/C100776233","wikidata":"https://www.wikidata.org/wiki/Q2532492","display_name":"Bridge (graph theory)","level":2,"score":0.3140000104904175},{"id":"https://openalex.org/C2780165032","wikidata":"https://www.wikidata.org/wiki/Q16869822","display_name":"Energy consumption","level":2,"score":0.31130000948905945},{"id":"https://openalex.org/C186370098","wikidata":"https://www.wikidata.org/wiki/Q442787","display_name":"Energy (signal processing)","level":2,"score":0.28540000319480896},{"id":"https://openalex.org/C7149132","wikidata":"https://www.wikidata.org/wiki/Q1377840","display_name":"Forgetting","level":2,"score":0.27639999985694885},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.2612000107765198},{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.260699987411499},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.2583000063896179},{"id":"https://openalex.org/C61326573","wikidata":"https://www.wikidata.org/wiki/Q1496376","display_name":"Gaussian process","level":3,"score":0.257999986410141},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.25679999589920044},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.2547000050544739}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2603.03491","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2603.03491","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.03491","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,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:doi:10.48550/arxiv.2603.03491","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"sustainable_development_goals":[{"display_name":"Affordable and clean energy","score":0.931912362575531,"id":"https://metadata.un.org/sdg/7"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Compute-in-memory":[0],"(CiM)":[1],"architectures":[2],"promise":[3],"significant":[4],"improvements":[5],"in":[6,48,97,204],"energy":[7],"efficiency":[8,138],"and":[9,38,45,61,72,94,108,165,185,206],"throughput":[10],"for":[11,182,201],"deep":[12],"neural":[13,59,153,193],"network":[14],"acceleration":[15],"by":[16,151],"alleviating":[17],"the":[18,54,177,199],"von":[19],"Neumann":[20],"bottleneck.":[21],"However,":[22],"their":[23,202],"reliance":[24],"on":[25,112,195],"emerging":[26,196],"non-volatile":[27],"memory":[28,197],"devices":[29],"introduces":[30],"device-level":[31],"non-idealities-such":[32],"as":[33],"write":[34],"variability,":[35],"conductance":[36],"drift,":[37],"stochastic":[39],"noise-that":[40],"fundamentally":[41],"challenge":[42],"reliability,":[43],"predictability,":[44],"safety,":[46],"especially":[47],"safety-critical":[49,98],"applications.":[50],"This":[51],"talk":[52],"examines":[53],"reliability":[55,134],"limits":[56],"of":[57,65,179],"CiM-based":[58],"accelerators":[60,184],"presents":[62],"a":[63,102,118,143,187],"series":[64],"techniques":[66],"that":[67,82,122,146],"bridge":[68],"device":[69,85],"physics,":[70],"architecture,":[71],"learning":[73],"algorithms":[74],"to":[75,89],"address":[76],"these":[77,174],"challenges.":[78],"We":[79],"first":[80],"demonstrate":[81],"even":[83],"small":[84],"variations":[86],"can":[87],"lead":[88],"disproportionately":[90],"large":[91],"accuracy":[92],"degradation":[93],"catastrophic":[95],"failures":[96],"inference":[99,194],"workloads,":[100],"revealing":[101],"critical":[103],"gap":[104],"between":[105],"average-case":[106],"evaluations":[107],"worst-case":[109,149],"behavior.":[110],"Building":[111],"this":[113],"insight,":[114],"we":[115,141],"introduce":[116],"SWIM,":[117],"selective":[119],"write-verify":[120],"mechanism":[121],"strategically":[123],"applies":[124],"verification":[125],"only":[126],"where":[127],"it":[128],"is":[129],"most":[130],"impactful,":[131],"significantly":[132],"improving":[133],"while":[135],"maintaining":[136],"CiM's":[137],"advantages.":[139],"Finally,":[140],"explore":[142],"learning-centric":[144],"solution":[145],"improves":[147],"realistic":[148],"performance":[150],"training":[152,160],"networks":[154],"with":[155,162],"right-censored":[156],"Gaussian":[157],"noise,":[158],"aligning":[159],"assumptions":[161],"hardware-induced":[163],"variability":[164],"enabling":[166],"robust":[167],"deployment":[168],"without":[169],"excessive":[170],"hardware":[171],"overhead.":[172],"Together,":[173],"works":[175],"highlight":[176],"necessity":[178],"cross-layer":[180],"co-design":[181],"CiM":[183],"provide":[186],"principled":[188],"path":[189],"toward":[190],"dependable,":[191],"efficient":[192],"technologies-paving":[198],"way":[200],"adoption":[203],"safety-":[205],"reliability-critical":[207],"systems.":[208]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-03-06T00:00:00"}