{"id":"https://openalex.org/W2963034893","doi":null,"title":"Efficient softmax approximation for GPUs","display_name":"Efficient softmax approximation for GPUs","publication_year":2017,"publication_date":"2017-08-06","ids":{"openalex":"https://openalex.org/W2963034893","mag":"2963034893"},"language":"en","primary_location":{"id":"mag:2963034893","doi":null,"is_oa":false,"landing_page_url":"http://proceedings.mlr.press/v70/grave17a/grave17a.pdf","pdf_url":null,"source":{"id":"https://openalex.org/S4306419644","display_name":"International Conference on Machine Learning","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":"International Conference on Machine Learning","raw_type":null},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":[],"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/A5069316249","display_name":"\u00c9douard Grave","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"\u00c9douard Grave","raw_affiliation_strings":["Facebook AI Research#TAB#"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Facebook AI Research#TAB#","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5107859338","display_name":"Armand Joulin","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Armand Joulin","raw_affiliation_strings":["Facebook AI Research#TAB#"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Facebook AI Research#TAB#","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5008590331","display_name":"Moustapha Ciss\u00e9","orcid":"https://orcid.org/0000-0003-2400-2492","observed_orcids":["https://orcid.org/0000-0003-2400-2492"]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Moustapha Ciss\u00e9","raw_affiliation_strings":["Facebook AI Research#TAB#"],"raw_orcid":"https://orcid.org/0000-0003-2400-2492","affiliations":[{"raw_affiliation_string":"Facebook AI Research#TAB#","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5111802678","display_name":"Herv\u00e9 Je\u01f5ou","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Herv\u00e9 Je\u01f5ou","raw_affiliation_strings":["Facebook AI Research#TAB#"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Facebook AI Research#TAB#","institution_ids":["https://openalex.org/I4210114444"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I4210114444"],"apc_list":null,"apc_paid":null,"fwci":8.3691,"has_fulltext":false,"cited_by_count":101,"citation_normalized_percentile":{"value":0.98250327,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":88,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"1302","last_page":"1310"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.6039999723434448,"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/T10028","display_name":"Topic Modeling","score":0.6039999723434448,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.23819999396800995,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.06560000032186508,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/vocabulary-size-reduction","display_name":"vocabulary size reduction","score":0.6669999957084656},{"id":"https://openalex.org/keywords/neural-language-models","display_name":"neural language models","score":0.6930000185966492},{"id":"https://openalex.org/keywords/gpu-acceleration","display_name":"GPU acceleration","score":0.7689999938011169},{"id":"https://openalex.org/keywords/training-efficiency","display_name":"training efficiency","score":0.6200000047683716},{"id":"https://openalex.org/keywords/large-vocabulary","display_name":"large vocabulary","score":0.6370000243186951}],"concepts":[{"id":"https://openalex.org/C188441871","wikidata":"https://www.wikidata.org/wiki/Q7554146","display_name":"Softmax function","level":3,"score":0.9670206904411316},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8010696172714233},{"id":"https://openalex.org/C19768560","wikidata":"https://www.wikidata.org/wiki/Q320727","display_name":"Dependency (UML)","level":2,"score":0.6420581936836243},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.5837069749832153},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5691214799880981},{"id":"https://openalex.org/C2777601683","wikidata":"https://www.wikidata.org/wiki/Q6499736","display_name":"Vocabulary","level":2,"score":0.5636730194091797},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.514164924621582},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4603132903575897},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4353092610836029},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.40816134214401245},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3952556848526001},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.11028420925140381},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.08440467715263367},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","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/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"mag:2963034893","doi":null,"is_oa":false,"landing_page_url":"http://proceedings.mlr.press/v70/grave17a/grave17a.pdf","pdf_url":null,"source":{"id":"https://openalex.org/S4306419644","display_name":"International Conference on Machine Learning","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":"International Conference on Machine Learning","raw_type":null}],"best_oa_location":null,"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":45,"referenced_works":["https://openalex.org/W22168010","https://openalex.org/W36903255","https://openalex.org/W179875071","https://openalex.org/W932413789","https://openalex.org/W1558797106","https://openalex.org/W1614298861","https://openalex.org/W1800356822","https://openalex.org/W1924770834","https://openalex.org/W1934041838","https://openalex.org/W1965154800","https://openalex.org/W1966812932","https://openalex.org/W1970689298","https://openalex.org/W1999965501","https://openalex.org/W2026383756","https://openalex.org/W2057653135","https://openalex.org/W2064675550","https://openalex.org/W2079656678","https://openalex.org/W2100664567","https://openalex.org/W2100714283","https://openalex.org/W2110485445","https://openalex.org/W2111305191","https://openalex.org/W2118776487","https://openalex.org/W2121227244","https://openalex.org/W2127836646","https://openalex.org/W2130942839","https://openalex.org/W2131462252","https://openalex.org/W2132339004","https://openalex.org/W2134237567","https://openalex.org/W2143612262","https://openalex.org/W2143719855","https://openalex.org/W2146502635","https://openalex.org/W2150355110","https://openalex.org/W2152790380","https://openalex.org/W2152808281","https://openalex.org/W2160815625","https://openalex.org/W2171928131","https://openalex.org/W2185726469","https://openalex.org/W2250379827","https://openalex.org/W2251682575","https://openalex.org/W2259472270","https://openalex.org/W2293997542","https://openalex.org/W2295800168","https://openalex.org/W2949947856","https://openalex.org/W2951714314","https://openalex.org/W2963603213"],"related_works":["https://openalex.org/W2750846973","https://openalex.org/W2742102274","https://openalex.org/W2898744876","https://openalex.org/W2739909019","https://openalex.org/W2808168001","https://openalex.org/W2152808281","https://openalex.org/W2741884846","https://openalex.org/W2963932686","https://openalex.org/W4285101301","https://openalex.org/W2175585630"],"abstract_inverted_index":{"We":[0],"propose":[1],"an":[2,102],"approximate":[3],"strategy":[4],"to":[5,36,105],"efficiently":[6],"train":[7],"neural":[8],"network":[9],"based":[10],"language":[11],"models":[12],"over":[13,97],"very":[14],"large":[15,93],"vocabularies.":[16],"Our":[17,47,73],"approach,":[18],"called":[19],"adaptive":[20],"softmax,":[21],"circumvents":[22],"the":[23,27,32,42,51,56,108],"linear":[24],"dependency":[25],"on":[26,77],"vocabulary":[28],"size":[29],"by":[30,54],"exploiting":[31,55],"unbalanced":[33],"word":[34],"distribution":[35],"form":[37],"clusters":[38],"that":[39,88,106],"explicitly":[40],"minimize":[41],"expectation":[43],"of":[44,58,107,113],"computation":[45],"time.":[46],"approach":[48,90],"further":[49],"reduces":[50],"computational":[52],"time":[53],"specificities":[57],"modern":[59],"architectures":[60],"and":[61,83],"matrix-matrix":[62],"vector":[63],"operations,":[64],"making":[65],"it":[66],"particularly":[67],"suited":[68],"for":[69],"graphical":[70],"processing":[71],"units.":[72],"experiments":[74],"carried":[75],"out":[76],"standard":[78,98],"benchmarks,":[79],"such":[80],"as":[81],"EuroParl":[82],"One":[84],"Billion":[85],"Word,":[86],"show":[87],"our":[89,114],"brings":[91],"a":[92],"gain":[94],"in":[95],"efficiency":[96],"approximations":[99],"while":[100],"achieving":[101],"accuracy":[103],"close":[104],"full":[109],"softmax.":[110],"The":[111],"code":[112],"method":[115],"is":[116],"available":[117],"at":[118],"https://github.com/facebookresearch/adaptive-softmax.":[119]},"counts_by_year":[{"year":2023,"cited_by_count":1},{"year":2022,"cited_by_count":2},{"year":2021,"cited_by_count":22},{"year":2020,"cited_by_count":29},{"year":2019,"cited_by_count":30},{"year":2018,"cited_by_count":15},{"year":2017,"cited_by_count":2}],"updated_date":"2026-11-04T10:54:05.112534","created_date":"2019-07-30T00:00:00"}