{"id":"https://openalex.org/W3132041002","doi":null,"title":"Linear Transformers Are Secretly Fast Weight Memory Systems","display_name":"Linear Transformers Are Secretly Fast Weight Memory Systems","publication_year":2021,"publication_date":"2021-02-22","ids":{"openalex":"https://openalex.org/W3132041002","mag":"3132041002"},"language":"en","primary_location":{"id":"mag:3132041002","doi":null,"is_oa":true,"landing_page_url":"http://export.arxiv.org/pdf/2102.11174","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":"http://export.arxiv.org/pdf/2102.11174","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5019167028","display_name":"Imanol Schlag","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Imanol Schlag","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5002810304","display_name":"Kazuki Irie","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kazuki Irie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5071172037","display_name":"J\u00fcrgen Schmidhuber","orcid":"https://orcid.org/0000-0002-1468-6758","observed_orcids":["https://orcid.org/0000-0002-1468-6758"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"J\u00fcrgen Schmidhuber","raw_affiliation_strings":[],"raw_orcid":null,"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":14,"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/T10028","display_name":"Topic Modeling","score":0.8823000192642212,"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":0.8823000192642212,"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":0.028699999675154686,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.008999999612569809,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/softmax-attention","display_name":"softmax attention","score":0.703000009059906},{"id":"https://openalex.org/keywords/memory-capacity-limits","display_name":"memory capacity limits","score":0.703000009059906},{"id":"https://openalex.org/keywords/linear-attention","display_name":"linear attention","score":0.3370000123977661},{"id":"https://openalex.org/keywords/kernel-function-design","display_name":"kernel function design","score":0.6869999766349792},{"id":"https://openalex.org/keywords/machine-translation","display_name":"machine translation","score":0.9869999885559082},{"id":"https://openalex.org/keywords/language-modeling","display_name":"language modeling","score":0.9890000224113464}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7571871280670166},{"id":"https://openalex.org/C188441871","wikidata":"https://www.wikidata.org/wiki/Q7554146","display_name":"Softmax function","level":3,"score":0.7472302913665771},{"id":"https://openalex.org/C2780069185","wikidata":"https://www.wikidata.org/wiki/Q7977945","display_name":"Equivalence (formal languages)","level":2,"score":0.5293111205101013},{"id":"https://openalex.org/C2988012377","wikidata":"https://www.wikidata.org/wiki/Q29966452","display_name":"Functional equivalence","level":2,"score":0.4996640682220459},{"id":"https://openalex.org/C2776372474","wikidata":"https://www.wikidata.org/wiki/Q508291","display_name":"Simplicity","level":2,"score":0.4929358959197998},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.48868897557258606},{"id":"https://openalex.org/C203005215","wikidata":"https://www.wikidata.org/wiki/Q79798","display_name":"Machine translation","level":2,"score":0.46251311898231506},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.41099774837493896},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3887619376182556},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.12153658270835876},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.08044883608818054},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C118615104","wikidata":"https://www.wikidata.org/wiki/Q121416","display_name":"Discrete mathematics","level":1,"score":0.0},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"mag:3132041002","doi":null,"is_oa":true,"landing_page_url":"http://export.arxiv.org/pdf/2102.11174","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:3132041002","doi":null,"is_oa":true,"landing_page_url":"http://export.arxiv.org/pdf/2102.11174","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":44,"referenced_works":["https://openalex.org/W185620388","https://openalex.org/W1525859397","https://openalex.org/W1810943226","https://openalex.org/W2013494846","https://openalex.org/W2016760105","https://openalex.org/W2024585065","https://openalex.org/W2057529293","https://openalex.org/W2089217417","https://openalex.org/W2095425517","https://openalex.org/W2095705004","https://openalex.org/W2101105183","https://openalex.org/W2128084896","https://openalex.org/W2129217160","https://openalex.org/W2176412452","https://openalex.org/W2413794162","https://openalex.org/W2525332836","https://openalex.org/W2535697732","https://openalex.org/W2587333550","https://openalex.org/W2597655663","https://openalex.org/W2626778328","https://openalex.org/W2760103357","https://openalex.org/W2795851676","https://openalex.org/W2806311723","https://openalex.org/W2883582441","https://openalex.org/W2894175714","https://openalex.org/W2943845043","https://openalex.org/W2950159625","https://openalex.org/W2952509486","https://openalex.org/W2962784628","https://openalex.org/W2962940432","https://openalex.org/W2963267363","https://openalex.org/W2963341956","https://openalex.org/W2964110616","https://openalex.org/W2964121744","https://openalex.org/W2964138017","https://openalex.org/W2970971581","https://openalex.org/W3007776460","https://openalex.org/W3034573343","https://openalex.org/W3099920963","https://openalex.org/W3113055895","https://openalex.org/W3116353560","https://openalex.org/W3120633509","https://openalex.org/W3120885796","https://openalex.org/W3125056032"],"related_works":["https://openalex.org/W3166702123","https://openalex.org/W7160998411","https://openalex.org/W2521593519","https://openalex.org/W7118324814","https://openalex.org/W7204843739","https://openalex.org/W4401157551","https://openalex.org/W4413298366","https://openalex.org/W7160835469","https://openalex.org/W3033529678","https://openalex.org/W7161839096"],"abstract_inverted_index":{"We":[0,76,90],"show":[1],"the":[2,14,46,66,109],"formal":[3],"equivalence":[4],"of":[5,26,38,48,111],"linearised":[6,28],"self-attention":[7],"mechanisms":[8],"and":[9,50,88,103],"fast":[10,39,60],"weight":[11,40],"memories":[12],"from":[13],"early":[15],"'90s.":[16],"From":[17],"this":[18],"observation":[19],"we":[20,62],"infer":[21],"a":[22,35,79],"memory":[23,41,49],"capacity":[24],"limitation":[25],"recent":[27],"softmax":[29],"attention":[30],"variants.":[31],"With":[32],"finite":[33],"memory,":[34],"desirable":[36],"behaviour":[37],"models":[42],"is":[43],"to":[44,64,83],"manipulate":[45],"contents":[47],"dynamically":[51],"interact":[52],"with":[53,69],"it.":[54],"Inspired":[55],"by":[56],"previous":[57],"work":[58],"on":[59,93],"weights,":[61],"propose":[63,78],"replace":[65],"update":[67],"rule":[68,72],"an":[70],"alternative":[71],"yielding":[73],"such":[74],"behaviour.":[75],"also":[77],"new":[80],"kernel":[81],"function":[82],"linearise":[84],"attention,":[85],"balancing":[86],"simplicity":[87],"effectiveness.":[89],"conduct":[91],"experiments":[92],"synthetic":[94],"retrieval":[95],"problems":[96],"as":[97,99],"well":[98],"standard":[100],"machine":[101],"translation":[102],"language":[104],"modelling":[105],"tasks":[106],"which":[107],"demonstrate":[108],"benefits":[110],"our":[112],"methods.":[113]},"counts_by_year":[{"year":2024,"cited_by_count":1},{"year":2023,"cited_by_count":1},{"year":2022,"cited_by_count":5},{"year":2021,"cited_by_count":5},{"year":2020,"cited_by_count":2}],"updated_date":"2026-11-03T10:50:19.383937","created_date":"2021-03-01T00:00:00"}