{"id":"https://openalex.org/W2554844166","doi":null,"title":"ZipML: An End-to-end Bitwise Framework for Dense Generalized Linear Models.","display_name":"ZipML: An End-to-end Bitwise Framework for Dense Generalized Linear Models.","publication_year":2016,"publication_date":"2016-11-16","ids":{"openalex":"https://openalex.org/W2554844166","mag":"2554844166"},"language":"en","primary_location":{"id":"mag:2554844166","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/1611.05402","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/abs/1611.05402","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101561475","display_name":"Hantian Zhang","orcid":"https://orcid.org/0000-0002-9862-8773","observed_orcids":["https://orcid.org/0000-0002-9862-8773"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hantian Zhang","raw_affiliation_strings":[],"raw_orcid":"https://orcid.org/0000-0002-9862-8773","affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5012725714","display_name":"Kaan Kara","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kaan Kara","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101439736","display_name":"Jerry Zheng Li","orcid":"https://orcid.org/0009-0003-5035-3728","observed_orcids":["https://orcid.org/0009-0003-5035-3728"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jerry Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5083822059","display_name":"Dan Alistarh","orcid":"https://orcid.org/0000-0003-3650-940X","observed_orcids":["https://orcid.org/0000-0003-3650-940X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dan Alistarh","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100396318","display_name":"Ji Liu","orcid":"https://orcid.org/0000-0002-0240-8881","observed_orcids":["https://orcid.org/0000-0002-0240-8881"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ji Liu","raw_affiliation_strings":[],"raw_orcid":"https://orcid.org/0000-0002-0240-8881","affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5004750246","display_name":"Ce Zhang","orcid":"https://orcid.org/0000-0002-8105-7505","observed_orcids":["https://orcid.org/0000-0002-8105-7505","https://orcid.org/0000-0001-5100-3584","https://orcid.org/0000-0001-6789-0130"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ce Zhang","raw_affiliation_strings":[],"raw_orcid":"https://orcid.org/0000-0001-5100-3584","affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":10,"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":0.8779000043869019,"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":0.8779000043869019,"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/T11697","display_name":"Numerical Methods and Algorithms","score":0.03480000048875809,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.031199999153614044,"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/low-precision-training","display_name":"low-precision training","score":0.8159999847412109},{"id":"https://openalex.org/keywords/generalized-linear-models","display_name":"generalized linear models","score":0.7110000252723694},{"id":"https://openalex.org/keywords/data-quantization","display_name":"data quantization","score":0.5199999809265137},{"id":"https://openalex.org/keywords/gradient-quantization","display_name":"gradient quantization","score":0.8349999785423279},{"id":"https://openalex.org/keywords/model-quantization","display_name":"model quantization","score":0.8069999814033508},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"stochastic gradient descent","score":0.7540000081062317},{"id":"https://openalex.org/keywords/bandwidth-compression","display_name":"bandwidth reduction","score":0.5979999899864197},{"id":"https://openalex.org/keywords/linear-regression","display_name":"linear regression","score":0.8650000095367432},{"id":"https://openalex.org/keywords/least-squares-svm","display_name":"least squares SVM","score":0.7789999842643738},{"id":"https://openalex.org/keywords/convergence-robustness","display_name":"convergence robustness","score":0.6480000019073486}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6884152293205261},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.5878589749336243},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5581665635108948},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.5336967706680298},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.5071592330932617},{"id":"https://openalex.org/C74296488","wikidata":"https://www.wikidata.org/wiki/Q2527392","display_name":"End-to-end principle","level":2,"score":0.4992959499359131},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.48551177978515625},{"id":"https://openalex.org/C134765980","wikidata":"https://www.wikidata.org/wiki/Q879126","display_name":"Bitwise operation","level":2,"score":0.476418137550354},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.4739679992198944},{"id":"https://openalex.org/C58489278","wikidata":"https://www.wikidata.org/wiki/Q1172284","display_name":"Data set","level":2,"score":0.41720911860466003},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.29436156153678894},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.09387189149856567},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0},{"id":"https://openalex.org/C50522688","wikidata":"https://www.wikidata.org/wiki/Q189833","display_name":"Economic growth","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/C94625758","wikidata":"https://www.wikidata.org/wiki/Q7163","display_name":"Politics","level":2,"score":0.0},{"id":"https://openalex.org/C199539241","wikidata":"https://www.wikidata.org/wiki/Q7748","display_name":"Law","level":1,"score":0.0},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","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},{"id":"https://openalex.org/C17744445","wikidata":"https://www.wikidata.org/wiki/Q36442","display_name":"Political science","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"mag:2554844166","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/1611.05402","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null}],"best_oa_location":{"id":"mag:2554844166","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/1611.05402","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2598713166","https://openalex.org/W2741269719","https://openalex.org/W2923673691","https://openalex.org/W4283322265","https://openalex.org/W2791110811","https://openalex.org/W7133190283","https://openalex.org/W3129739837","https://openalex.org/W3028224987","https://openalex.org/W3106325597","https://openalex.org/W4280575426"],"abstract_inverted_index":{"We":[0,176,257],"present":[1],"ZipML,":[2,15],"the":[3,43,57,101,118,128,130,160,189,218,247],"first":[4,152],"framework":[5],"for":[6,21,231],"training":[7,59,222],"dense":[8],"generalized":[9],"linear\nmodels":[10],"using":[11,28,94],"end-to-end":[12,76],"low-precision":[13,95],"representation--in":[14],"all":[16,140],"movements\nof":[17],"data,":[18,252],"including":[19],"those":[20],"input":[22,44],"samples,":[23,168],"model,":[24,253],"and":[25,50,98,162,169,197,236,254],"gradients,":[26],"are\nrepresented":[27],"as":[29,31,187,199,201,213],"little":[30,200],"two":[32,178],"bits":[33],"per":[34,203],"component.":[35,204],"Within":[36],"our":[37],"framework,":[38],"we\nhave":[39],"successfully":[40],"compressed,":[41],"separately,":[42],"data":[45,141,180,191,210],"by":[46,52,224,261],"16x,":[47],"gradient":[48,97,105,255],"by\n16x,":[49],"model":[51,99,161],"16x":[53],"while":[54,156],"still":[55],"getting":[56],"same":[58],"result.":[60,175],"Even":[61],"for\nthe":[62],"most":[63],"challenging":[64],"datasets,":[65],"we":[66,151,216],"find":[67],"that":[68],"robust":[69],"convergence":[70],"can":[71],"be":[72,121],"ensured\nusing":[73],"only":[74],"an":[75],"8-bit":[77],"representation":[78,82,192],"or":[79],"a":[80,111,137,173,206,262],"6-bit":[81],"if":[83,208],"only\nsamples":[84],"are":[85],"quantized.":[86],"Our":[87,107,228],"work":[88,266],"builds":[89],"on":[90,93],"previous":[91],"research":[92],"representations\nfor":[96],"in":[100,188,194],"context":[102],"of":[103,114,264],"stochastic":[104],"descent.":[106],"main\ntechnical":[108],"contribution":[109],"is":[110,164,211],"new":[112,179],"set":[113],"techniques":[115],"which":[116,182],"allow":[117],"training\nsamples":[119],"to":[120,136,172,184,226,246,250],"processed":[122],"with":[123],"low":[124,146],"precision,":[125],"without":[126],"affecting":[127],"convergence\nof":[129],"algorithm.":[131],"In":[132,149],"turn,":[133],"this":[134],"leads":[135,171],"system":[138],"where":[139],"items":[142],"move":[143],"in\na":[144],"quantized,":[145],"precision":[147],"format.":[148],"particular,":[150],"establish":[153],"that\nrandomized":[154],"rounding,":[155],"sufficient":[157],"when":[158,166],"quantizing":[159,167],"the\ngradients,":[163],"biased":[165],"thus":[170],"different\ntraining":[174],"propose":[177],"representations":[181],"converge":[183],"the\nsame":[185],"solution":[186],"original":[190],"both":[193],"theory":[195],"and\nempirically":[196],"require":[198],"2-bits":[202],"As":[205],"result,":[207],"the\noriginal":[209],"stored":[212],"32-bit":[214],"floats,":[215],"decrease":[217],"bandwidth":[219],"footprint\nfor":[220],"each":[221],"iteration":[223],"up":[225],"16x.":[227],"results":[229],"hold":[230],"models":[232],"such":[233],"as\nlinear":[234],"regression":[235],"least":[237],"squares":[238],"SVM.":[239],"ZipML":[240],"raises":[241],"interesting":[242],"theoretical":[243],"questions":[244],"related":[245],"robustness":[248],"of\nSGD":[249],"approximate":[251],"representations.":[256],"conclude":[258],"this\nworking":[259],"paper":[260],"description":[263],"ongoing":[265],"extending":[267],"these":[268],"preliminary\nresults.":[269]},"counts_by_year":[{"year":2020,"cited_by_count":2},{"year":2018,"cited_by_count":6},{"year":2017,"cited_by_count":2}],"updated_date":"2026-10-17T09:40:24.721904","created_date":"2016-11-30T00:00:00"}