{"id":"https://openalex.org/W4404088433","doi":"https://doi.org/10.1109/jsait.2026.3688610","title":"On the Convergence Theory of Pipeline Gradient-Based Analog In-Memory Training","display_name":"On the Convergence Theory of Pipeline Gradient-Based Analog In-Memory Training","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W4404088433","doi":"https://doi.org/10.1109/jsait.2026.3688610"},"language":"en","primary_location":{"id":"doi:10.1109/jsait.2026.3688610","is_oa":false,"landing_page_url":"https://doi.org/10.1109/jsait.2026.3688610","pdf_url":null,"source":{"id":"https://openalex.org/S4210211895","display_name":"IEEE Journal on Selected Areas in Information Theory","issn_l":"2641-8770","issn":["2641-8770"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Journal on Selected Areas in Information Theory","raw_type":"journal-article"},"type":"preprint","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2410.15155","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5074757613","display_name":"Zhaoxian Wu","orcid":"https://orcid.org/0000-0001-6724-6134"},"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":"Zhaoxian Wu","raw_affiliation_strings":["Cornell University, New York, NY, USA"],"raw_orcid":"https://orcid.org/0000-0001-6724-6134","affiliations":[{"raw_affiliation_string":"Cornell University, New York, NY, USA","institution_ids":["https://openalex.org/I205783295"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080123409","display_name":"Quan Xiao","orcid":"https://orcid.org/0000-0002-9565-3324"},"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":"Quan Xiao","raw_affiliation_strings":["Cornell University, New York, NY, USA"],"raw_orcid":"https://orcid.org/0009-0008-8492-0037","affiliations":[{"raw_affiliation_string":"Cornell University, New York, NY, USA","institution_ids":["https://openalex.org/I205783295"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073094974","display_name":"Tayfun Gokmen","orcid":"https://orcid.org/0000-0002-5677-1723"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tayfun Gokmen","raw_affiliation_strings":["IBM Research, Yorktown Heights, NY, USA"],"raw_orcid":"https://orcid.org/0000-0002-5677-1723","affiliations":[{"raw_affiliation_string":"IBM Research, Yorktown Heights, NY, USA","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086723230","display_name":"Hsinyu Tsai","orcid":"https://orcid.org/0000-0002-3971-097X"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hsinyu Tsai","raw_affiliation_strings":["IBM Research, Yorktown Heights, NY, USA"],"raw_orcid":"https://orcid.org/0000-0002-3971-097X","affiliations":[{"raw_affiliation_string":"IBM Research, Yorktown Heights, NY, USA","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5034513925","display_name":"Kaoutar El Maghraoui","orcid":"https://orcid.org/0000-0002-1967-8749"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kaoutar El Maghraoui","raw_affiliation_strings":["IBM Research, Yorktown Heights, NY, USA"],"raw_orcid":"https://orcid.org/0000-0002-1967-8749","affiliations":[{"raw_affiliation_string":"IBM Research, Yorktown Heights, NY, USA","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100783476","display_name":"Tianyi Chen","orcid":"https://orcid.org/0000-0003-3477-1439"},"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":"Tianyi Chen","raw_affiliation_strings":["Cornell University, New York, NY, USA"],"raw_orcid":"https://orcid.org/0000-0001-8893-3486","affiliations":[{"raw_affiliation_string":"Cornell University, New York, NY, USA","institution_ids":["https://openalex.org/I205783295"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":6,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.00099194,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"7","issue":null,"first_page":"279","last_page":"289"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9854999780654907,"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":0.9854999780654907,"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/T12676","display_name":"Machine Learning and ELM","score":0.9846000075340271,"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/T10320","display_name":"Neural Networks and Applications","score":0.9666000008583069,"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/pipeline","display_name":"Pipeline (software)","score":0.8271363973617554},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.7204185724258423},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5183635354042053},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3262110650539398},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.13090753555297852},{"id":"https://openalex.org/keywords/physics","display_name":"Physics","score":0.08217570185661316},{"id":"https://openalex.org/keywords/meteorology","display_name":"Meteorology","score":0.07737931609153748}],"concepts":[{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.8271363973617554},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.7204185724258423},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5183635354042053},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3262110650539398},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.13090753555297852},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.08217570185661316},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.07737931609153748}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.1109/jsait.2026.3688610","is_oa":false,"landing_page_url":"https://doi.org/10.1109/jsait.2026.3688610","pdf_url":null,"source":{"id":"https://openalex.org/S4210211895","display_name":"IEEE Journal on Selected Areas in Information Theory","issn_l":"2641-8770","issn":["2641-8770"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Journal on Selected Areas in Information Theory","raw_type":"journal-article"},{"id":"pmh:oai:arXiv.org:2410.15155","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2410.15155","pdf_url":"https://arxiv.org/pdf/2410.15155","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":null},{"id":"pmh:oai:arXiv.org:2410.15155","is_oa":true,"landing_page_url":"https://arxiv.org/abs/2410.15155","pdf_url":"https://arxiv.org/pdf/2410.15155","source":{"id":"https://openalex.org/S4393918464","display_name":"ArXiv.org","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2410.15155","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2410.15155","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"article"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2410.15155","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2410.15155","pdf_url":"https://arxiv.org/pdf/2410.15155","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":null},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1686446952","display_name":"Representation Learning via Variational Mean Field Theory","funder_award_id":"2401297","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G4794017400","display_name":"CAREER: Distributed Intelligence in Future Wireless Networks: From System-Aware Learning to System-Learning Co-designs","funder_award_id":"2532349","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G6203303384","display_name":"Collaborative Research: BLoG: A Bi-Level Optimization Framework for Learning Over Graphs","funder_award_id":"2532653","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320307791","display_name":"Cisco Systems","ror":"https://ror.org/03yt1ez60"},{"id":"https://openalex.org/F4320309500","display_name":"Rensselaer Polytechnic Institute","ror":"https://ror.org/01rtyzb94"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W230091440","https://openalex.org/W2390279801","https://openalex.org/W2233261550","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2810751659"],"abstract_inverted_index":{"Aiming":[0],"to":[1,38],"accelerate":[2],"the":[3,72,87,93,116,128,137,148,157,162,174,182,187,199,217],"training":[4,88,129,207],"of":[5,74,96,113,119,130,165,189],"large":[6],"deep":[7],"neural":[8],"networks":[9],"(DNN)":[10],"in":[11,31,124,142],"an":[12,104],"energy-efficient":[13],"way,":[14],"analog":[15,121],"in-memory":[16],"computing":[17],"(AIMC)":[18],"emerges":[19],"as":[20],"a":[21],"solution":[22],"with":[23,103,177,195,216],"immense":[24],"potential.":[25],"AIMC":[26,50,68,101,114,206],"accelerator":[27],"keeps":[28],"model":[29],"weights":[30,144],"memory":[32,37],"without":[33],"moving":[34],"them":[35],"from":[36,209],"processors":[39],"during":[40,86],"training,":[41],"reducing":[42],"overhead":[43],"dramatically.":[44],"Despite":[45],"its":[46],"efficiency,":[47],"scaling":[48],"up":[49],"systems":[51],"presents":[52],"significant":[53],"challenges.":[54],"Since":[55],"weight":[56,125],"copying":[57],"is":[58,64,110],"expensive":[59],"and":[60,192],"inaccurate,":[61],"data":[62],"parallelism":[63,140],"less":[65],"efficient":[66],"on":[67,100,167],"accelerators.":[69],"It":[70,203],"necessitates":[71],"exploration":[73,112],"pipeline":[75,79,106,139,219],"parallelism,":[76,80],"particularly":[77],"asynchronous":[78,105,138,210],"which":[81,146,185],"utilizes":[82],"all":[83],"available":[84],"accelerators":[85],"process.":[89],"This":[90],"paper":[91,160],"examines":[92],"convergence":[94,163],"theory":[95],"stochastic":[97],"gradient":[98],"descent":[99],"hardware":[102,122],"(Analog-SGD-AP).":[107],"Although":[108],"there":[109],"empirical":[111],"accelerators,":[115],"theoretical":[117],"understanding":[118],"how":[120],"imperfections":[123],"updates":[126],"affect":[127],"multi-layer":[131,168],"DNN":[132,169],"models":[133],"remains":[134],"underexplored.":[135],"Furthermore,":[136],"results":[141],"stale":[143],"issues,":[145,184],"render":[147],"update":[149],"signals":[150],"no":[151],"longer":[152],"valid":[153],"gradients.":[154],"To":[155],"close":[156],"gap,":[158],"this":[159],"investigates":[161],"properties":[164],"Analog-SGD-AP":[166,175],"training.":[170],"We":[171],"show":[172],"that":[173,205],"converges":[176],"iteration":[178],"complexity":[179],"$O(\\varepsilon^{-2}+\\varepsilon^{-1})$":[180],"despite":[181],"aforementioned":[183],"matches":[186],"complexities":[188],"digital":[190],"SGD":[191,194],"Analog":[193],"synchronous":[196,218],"pipeline,":[197],"except":[198],"non-dominant":[200],"term":[201],"$O(\\varepsilon^{-1})$.":[202],"implies":[204],"benefits":[208],"pipelining":[211],"almost":[212],"for":[213],"free":[214],"compared":[215],"by":[220],"overlapping":[221],"computation.":[222]},"counts_by_year":[],"updated_date":"2026-06-22T08:00:12.763002","created_date":"2025-10-10T00:00:00"}