{"id":"https://openalex.org/W4388650694","doi":"https://doi.org/10.48550/arxiv.2311.05760","title":"Compressed and Sparse Models for Non-Convex Decentralized Learning","display_name":"Compressed and Sparse Models for Non-Convex Decentralized Learning","publication_year":2023,"publication_date":"2023-11-09","ids":{"openalex":"https://openalex.org/W4388650694","doi":"https://doi.org/10.48550/arxiv.2311.05760"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2311.05760","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2311.05760","pdf_url":"https://arxiv.org/pdf/2311.05760","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,"listed_in":[],"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":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2311.05760","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Campbell, Andrew","orcid":null,"observed_orcids":[]},"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":"Campbell, Andrew","raw_affiliation_strings":["Electrical and Computer Engineering Cornell University Ithaca , NY 14850"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Electrical and Computer Engineering Cornell University Ithaca , NY 14850","institution_ids":["https://openalex.org/I205783295"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100338967","display_name":"Hang Liu","orcid":"https://orcid.org/0000-0002-5246-8399","observed_orcids":["https://orcid.org/0000-0002-5246-8399"]},"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":"Liu, Hang","raw_affiliation_strings":["Electrical and Computer Engineering Cornell University Ithaca , NY 14850"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Electrical and Computer Engineering Cornell University Ithaca , NY 14850","institution_ids":["https://openalex.org/I205783295"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5084646801","display_name":"Leah Woldemariam","orcid":"https://orcid.org/0000-0002-6809-1071","observed_orcids":["https://orcid.org/0000-0002-6809-1071"]},"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":"Woldemariam, Leah","raw_affiliation_strings":["Electrical and Computer Engineering Cornell University Ithaca , NY 14850"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Electrical and Computer Engineering Cornell University Ithaca , NY 14850","institution_ids":["https://openalex.org/I205783295"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5029881017","display_name":"Anna Scaglione","orcid":"https://orcid.org/0000-0002-8892-3680","observed_orcids":["https://orcid.org/0000-0002-8892-3680"]},"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":"Scaglione, Anna","raw_affiliation_strings":["Electrical and Computer Engineering Cornell University Ithaca , NY 14850"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Electrical and Computer Engineering Cornell University Ithaca , NY 14850","institution_ids":["https://openalex.org/I205783295"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I205783295"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9624000191688538,"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"}},"topics":[{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9624000191688538,"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/T10249","display_name":"Distributed Control Multi-Agent Systems","score":0.020999999716877937,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems 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/T10764","display_name":"Privacy-Preserving Technologies in Data","score":0.007899999618530273,"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/decentralized-learning","display_name":"decentralized learning","score":0.828000009059906},{"id":"https://openalex.org/keywords/gradient-compression","display_name":"gradient compression","score":0.8389999866485596},{"id":"https://openalex.org/keywords/model-sparsification","display_name":"model sparsification","score":0.9760000109672546},{"id":"https://openalex.org/keywords/nonconvex-optimization","display_name":"non-convex optimization","score":0.9440000057220459},{"id":"https://openalex.org/keywords/l1-regularization","display_name":"L1 regularization","score":0.9210000038146973},{"id":"https://openalex.org/keywords/communication-efficiency","display_name":"communication efficiency","score":0.8209999799728394},{"id":"https://openalex.org/keywords/convergence-rate","display_name":"convergence rate","score":0.6259999871253967},{"id":"https://openalex.org/keywords/overparameterized-neural-networks","display_name":"overparameterized neural networks","score":0.7739999890327454}],"concepts":[{"id":"https://openalex.org/C70451592","wikidata":"https://www.wikidata.org/wiki/Q376493","display_name":"Dither","level":3,"score":0.6674603819847107},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.64762282371521},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.6043747663497925},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.5285405516624451},{"id":"https://openalex.org/C165464430","wikidata":"https://www.wikidata.org/wiki/Q1570441","display_name":"Parameterized complexity","level":2,"score":0.4989657402038574},{"id":"https://openalex.org/C57869625","wikidata":"https://www.wikidata.org/wiki/Q1783502","display_name":"Rate of convergence","level":3,"score":0.47037166357040405},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.4481738805770874},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.44004836678504944},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.43970930576324463},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3620700240135193},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3595714569091797},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3363453149795532},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.19077980518341064},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.08343416452407837},{"id":"https://openalex.org/C9083635","wikidata":"https://www.wikidata.org/wiki/Q2133535","display_name":"Noise shaping","level":2,"score":0.0},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.0},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.0},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2311.05760","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2311.05760","pdf_url":"https://arxiv.org/pdf/2311.05760","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,"listed_in":[],"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":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2311.05760","doi":"https://doi.org/10.48550/arxiv.2311.05760","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2311.05760","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,"listed_in":[],"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":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2311.05760","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2311.05760","pdf_url":"https://arxiv.org/pdf/2311.05760","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,"listed_in":[],"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":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4388650694.pdf","grobid_xml":"https://content.openalex.org/works/W4388650694.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W3119652720","https://openalex.org/W2982648258","https://openalex.org/W4390074148","https://openalex.org/W2963264832","https://openalex.org/W3047411835","https://openalex.org/W3101542044","https://openalex.org/W4387869894","https://openalex.org/W3024856659","https://openalex.org/W6967233052","https://openalex.org/W3127859550"],"abstract_inverted_index":{"Recent":[0],"research":[1],"highlights":[2],"frequent":[3],"model":[4,43,47],"communication":[5,78,134,151],"as":[6],"a":[7,32,58,88,100,125,129],"significant":[8],"bottleneck":[9],"to":[10,53,95,158],"the":[11,54,75,96,114,133,159],"efficiency":[12],"of":[13,79,91,116],"decentralized":[14,34,59],"machine":[15],"learning":[16,104],"(ML),":[17],"especially":[18],"for":[19,63,74,113,132],"large-scale":[20],"and":[21,56,71,103,145],"over-parameterized":[22],"neural":[23],"networks":[24],"(NNs).":[25],"To":[26],"address":[27],"this,":[28],"we":[29,123],"present":[30,57],"Malcom-PSGD,":[31],"novel":[33],"ML":[35],"algorithm":[36],"that":[37,85,147],"combines":[38],"gradient":[39,77],"compression":[40],"techniques":[41],"with":[42,93,137],"sparsification.":[44],"We":[45],"promote":[46],"sparsity":[48],"by":[49,110,153],"adding":[50],"$\\ell_1$":[51],"regularization":[52],"objective":[55],"proximal":[60],"SGD":[61,120],"method":[62,149],"training.":[64],"Our":[65,82],"approach":[66],"employs":[67],"vector":[68],"source":[69],"coding":[70],"dithering-based":[72],"quantization":[73],"compressed":[76,118],"sparsified":[80],"models.":[81],"analysis":[83],"demonstrates":[84],"Malcom-PSGD":[86],"achieves":[87],"convergence":[89,115],"rate":[90],"$\\mathcal{O}(1/\\sqrt{t})$":[92],"respect":[94],"iterations":[97],"$t$,":[98],"assuming":[99],"constant":[101],"consensus":[102],"rate.":[105],"This":[106],"result":[107],"is":[108],"supported":[109],"our":[111,142,148],"proof":[112],"non-convex":[117],"Proximal":[119],"methods.":[121],"Additionally,":[122],"conduct":[124],"bit":[126],"analysis,":[127],"providing":[128],"closed-form":[130],"expression":[131],"costs":[135,152],"associated":[136],"Malcom-PSGD.":[138],"Numerical":[139],"results":[140],"verify":[141],"theoretical":[143],"findings":[144],"demonstrate":[146],"reduces":[150],"approximately":[154],"$75\\%$":[155],"when":[156],"compared":[157],"state-of-the-art.":[160]},"counts_by_year":[],"updated_date":"2026-09-30T09:27:27.532063","created_date":"2023-11-14T00:00:00"}