{"id":"https://openalex.org/W2202715214","doi":"https://doi.org/10.1109/bigdata.2015.7363760","title":"Machine learning at the limit","display_name":"Machine learning at the limit","publication_year":2015,"publication_date":"2015-10-01","ids":{"openalex":"https://openalex.org/W2202715214","doi":"https://doi.org/10.1109/bigdata.2015.7363760","mag":"2202715214"},"language":"en","primary_location":{"id":"doi:10.1109/bigdata.2015.7363760","is_oa":false,"landing_page_url":"https://doi.org/10.1109/bigdata.2015.7363760","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2015 IEEE International Conference on Big Data (Big Data)","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/A5089723214","display_name":"John Canny","orcid":"https://orcid.org/0000-0002-7161-7927"},"institutions":[{"id":"https://openalex.org/I134446601","display_name":"Berkeley College","ror":"https://ror.org/02xewxa75","country_code":"US","type":"education","lineage":["https://openalex.org/I134446601"]},{"id":"https://openalex.org/I95457486","display_name":"University of California, Berkeley","ror":"https://ror.org/01an7q238","country_code":"US","type":"education","lineage":["https://openalex.org/I95457486"]},{"id":"https://openalex.org/I4210134091","display_name":"Yahoo (United States)","ror":"https://ror.org/040dkzz12","country_code":"US","type":"company","lineage":["https://openalex.org/I4210134091"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"John Canny","raw_affiliation_strings":["UC Berkeley, Berkeley, CA, USA","Yahoo Research, Sunnyvale, CA, USA"],"affiliations":[{"raw_affiliation_string":"UC Berkeley, Berkeley, CA, USA","institution_ids":["https://openalex.org/I134446601","https://openalex.org/I95457486"]},{"raw_affiliation_string":"Yahoo Research, Sunnyvale, CA, USA","institution_ids":["https://openalex.org/I4210134091"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5046520437","display_name":"Huasha Zhao","orcid":null},"institutions":[{"id":"https://openalex.org/I134446601","display_name":"Berkeley College","ror":"https://ror.org/02xewxa75","country_code":"US","type":"education","lineage":["https://openalex.org/I134446601"]},{"id":"https://openalex.org/I95457486","display_name":"University of California, Berkeley","ror":"https://ror.org/01an7q238","country_code":"US","type":"education","lineage":["https://openalex.org/I95457486"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Huasha Zhao","raw_affiliation_strings":["UC Berkeley, Berkeley, CA, USA"],"affiliations":[{"raw_affiliation_string":"UC Berkeley, Berkeley, CA, USA","institution_ids":["https://openalex.org/I134446601","https://openalex.org/I95457486"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039683044","display_name":"Bobby Jaros","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bobby Jaros","raw_affiliation_strings":["Dianping.com, Changning District, Shanghai, China"],"affiliations":[{"raw_affiliation_string":"Dianping.com, Changning District, Shanghai, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100359095","display_name":"Ye Chen","orcid":"https://orcid.org/0000-0002-1080-7671"},"institutions":[{"id":"https://openalex.org/I4210134091","display_name":"Yahoo (United States)","ror":"https://ror.org/040dkzz12","country_code":"US","type":"company","lineage":["https://openalex.org/I4210134091"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ye Chen","raw_affiliation_strings":["Yahoo Research, Sunnyvale, CA, USA"],"affiliations":[{"raw_affiliation_string":"Yahoo Research, Sunnyvale, CA, USA","institution_ids":["https://openalex.org/I4210134091"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5037616569","display_name":"Jiangchang Mao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiangchang Mao","raw_affiliation_strings":["Dianping.com, Changning District, Shanghai, China"],"affiliations":[{"raw_affiliation_string":"Dianping.com, Changning District, Shanghai, China","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5089723214"],"corresponding_institution_ids":["https://openalex.org/I134446601","https://openalex.org/I4210134091","https://openalex.org/I95457486"],"apc_list":null,"apc_paid":null,"fwci":2.9068,"has_fulltext":false,"cited_by_count":26,"citation_normalized_percentile":{"value":0.91087681,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":89,"max":98},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9998999834060669,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9998999834060669,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9987000226974487,"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.9970999956130981,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/computer-science","display_name":"Computer science","score":0.7971378564834595},{"id":"https://openalex.org/keywords/limit","display_name":"Limit (mathematics)","score":0.516667366027832},{"id":"https://openalex.org/keywords/node","display_name":"Node (physics)","score":0.48808348178863525},{"id":"https://openalex.org/keywords/reduction","display_name":"Reduction (mathematics)","score":0.4568513035774231},{"id":"https://openalex.org/keywords/transformation","display_name":"Transformation (genetics)","score":0.41972413659095764},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3321589529514313},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.32061147689819336},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.10576727986335754}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7971378564834595},{"id":"https://openalex.org/C151201525","wikidata":"https://www.wikidata.org/wiki/Q177239","display_name":"Limit (mathematics)","level":2,"score":0.516667366027832},{"id":"https://openalex.org/C62611344","wikidata":"https://www.wikidata.org/wiki/Q1062658","display_name":"Node (physics)","level":2,"score":0.48808348178863525},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.4568513035774231},{"id":"https://openalex.org/C204241405","wikidata":"https://www.wikidata.org/wiki/Q461499","display_name":"Transformation (genetics)","level":3,"score":0.41972413659095764},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3321589529514313},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.32061147689819336},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.10576727986335754},{"id":"https://openalex.org/C55493867","wikidata":"https://www.wikidata.org/wiki/Q7094","display_name":"Biochemistry","level":1,"score":0.0},{"id":"https://openalex.org/C66938386","wikidata":"https://www.wikidata.org/wiki/Q633538","display_name":"Structural engineering","level":1,"score":0.0},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.0},{"id":"https://openalex.org/C104317684","wikidata":"https://www.wikidata.org/wiki/Q7187","display_name":"Gene","level":2,"score":0.0},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/bigdata.2015.7363760","is_oa":false,"landing_page_url":"https://doi.org/10.1109/bigdata.2015.7363760","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2015 IEEE International Conference on Big Data (Big Data)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/7","display_name":"Affordable and clean energy","score":0.7200000286102295}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":23,"referenced_works":["https://openalex.org/W78077100","https://openalex.org/W1695021962","https://openalex.org/W1788180225","https://openalex.org/W1934084512","https://openalex.org/W1999702719","https://openalex.org/W2002555321","https://openalex.org/W2039548408","https://openalex.org/W2060393849","https://openalex.org/W2083842231","https://openalex.org/W2131975293","https://openalex.org/W2153579005","https://openalex.org/W2157462866","https://openalex.org/W2951113132","https://openalex.org/W3004511899","https://openalex.org/W4285719527","https://openalex.org/W4294170691","https://openalex.org/W6603201521","https://openalex.org/W6638233953","https://openalex.org/W6639805184","https://openalex.org/W6665801690","https://openalex.org/W6679815717","https://openalex.org/W6682691769","https://openalex.org/W6683124362"],"related_works":["https://openalex.org/W2353396818","https://openalex.org/W2370473919","https://openalex.org/W2532234348","https://openalex.org/W2762984199","https://openalex.org/W2353644209","https://openalex.org/W2390496461","https://openalex.org/W2389654532","https://openalex.org/W108084911","https://openalex.org/W2593663830","https://openalex.org/W2348743188"],"abstract_inverted_index":{"Many":[0],"systems":[1],"have":[2,71,190],"been":[3,17],"developed":[4,72],"for":[5,65,85,200,205],"machine":[6,48,148,207],"learning":[7,149],"at":[8],"scale.":[9],"Performance":[10],"has":[11,16,78],"steadily":[12],"improved,":[13],"but":[14],"there":[15],"relatively":[18,114],"little":[19],"work":[20,181],"on":[21,103,182,242,261],"explicitly":[22],"defining":[23],"or":[24,265],"approaching":[25],"the":[26,35,61,80,128,144,195,232],"limits":[27],"of":[28,37,105,117,147,239],"performance.":[29,168],"In":[30,50],"this":[31],"paper":[32],"we":[33,70,211,235],"describe":[34,179,212,236],"application":[36],"roofline":[38,51,68,138],"design,":[39,52,69],"an":[40,213],"approach":[41,159],"borrowed":[42],"from":[43,216],"computer":[44],"architecture,":[45],"to":[46,83,131,143,155,164,175,218,231],"large-scale":[47],"learning.":[49,208],"one":[53,90],"exposes":[54],"ALU,":[55],"memory,":[56],"and":[57,60,99],"network":[58,176,197],"limits,":[59],"constraints":[62],"they":[63],"imply":[64],"algorithms.":[66],"Using":[67,209],"a":[73,113,124,161,183,202,224],"system":[74],"called":[75,187],"BIDMach":[76],"which":[77,151],"demonstrated":[79,256],"highest":[81],"performance":[82,108,122],"date":[84],"many":[86],"ML":[87],"problems.":[88],"On":[89],"GPU-accelerated":[91],"node,":[92],"it":[93],"generally":[94],"outperforms":[95],"other":[96],"single-machine":[97],"toolkits":[98,101],"cluster":[100],"running":[102],"100s":[104],"nodes.":[106],"This":[107,158,221],"level":[109],"is":[110],"enabled":[111],"by":[112],"small":[115],"number":[116],"rooflined":[118,254],"matrix":[119,136],"primitives.":[120,177],"Such":[121],"implies":[123],"dramatic":[125],"reduction":[126],"in":[127,167,245],"energy":[129],"used":[130],"perform":[132],"these":[133,240],"calculations.":[134,220],"Beyond":[135],"kernels,":[137],"design":[139,146,170,255],"can":[140,171],"be":[141,173],"applied":[142,174],"end-to-end":[145],"algorithms":[150],"minimize":[152],"memory":[153],"usage":[154],"optimize":[156],"speed.":[157],"offers":[160],"further":[162],"2x":[163],"3x":[165],"gain":[166],"Roofline":[169],"also":[172],"We":[178,189],"recent":[180],"sparse":[184],"allreduce":[185],"primitive":[186,204],"Kylix.":[188],"shown":[191],"that":[192],"Kylix":[193],"approaches":[194],"practical":[196],"throughput":[198],"limit":[199],"allreduce,":[201],"basic":[203],"distributed":[206],"Kylix,":[210],"efficient":[214],"transformation":[215,222],"model-parallel":[217],"data-parallel":[219],"uses":[223],"secondary":[225],"storage":[226],"roofline,":[227],"with":[228],"similar":[229],"parameters":[230],"network.":[233],"Finally,":[234],"several":[237],"deployments":[238],"techniques":[241],"real-world":[243],"problems":[244],"two":[246],"large":[247],"internet":[248],"companies.":[249],"Once":[250],"again,":[251],"single":[252,263],"node":[253],"substantial":[257],"gains":[258],"over":[259],"alternatives":[260],"either":[262],"nodes":[264],"clusters.":[266]},"counts_by_year":[{"year":2025,"cited_by_count":1},{"year":2024,"cited_by_count":5},{"year":2023,"cited_by_count":5},{"year":2021,"cited_by_count":1},{"year":2020,"cited_by_count":1},{"year":2019,"cited_by_count":4},{"year":2018,"cited_by_count":3},{"year":2017,"cited_by_count":2},{"year":2016,"cited_by_count":4}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}