{"id":"https://openalex.org/W7212071069","doi":"https://doi.org/10.48550/arxiv.2609.09157","title":"Learning Length-Extrapolatable Recurrent Models","display_name":"Learning Length-Extrapolatable Recurrent Models","publication_year":2026,"publication_date":"2026-09-08","ids":{"openalex":"https://openalex.org/W7212071069","doi":"https://doi.org/10.48550/arxiv.2609.09157"},"language":null,"primary_location":{"id":"pmh:oai:arXiv.org:2609.09157","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.09157","pdf_url":"https://arxiv.org/pdf/2609.09157","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","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.09157","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5076386225","display_name":"Hanwen Jiang","orcid":"https://orcid.org/0000-0002-1523-8551","observed_orcids":["https://orcid.org/0000-0002-1523-8551"]},"institutions":[{"id":"https://openalex.org/I1306409833","display_name":"Adobe Systems (United States)","ror":"https://ror.org/059tvcg64","country_code":"US","type":"company","lineage":["https://openalex.org/I1306409833"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Jiang, Hanwen","raw_affiliation_strings":["Adobe Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Adobe Research","institution_ids":["https://openalex.org/I1306409833"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":["https://openalex.org/A5076386225"],"corresponding_institution_ids":["https://openalex.org/I1306409833"],"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/T10320","display_name":"Neural Networks and Applications","score":0.6304000020027161,"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/T10320","display_name":"Neural Networks and Applications","score":0.6304000020027161,"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/T10028","display_name":"Topic Modeling","score":0.14839999377727509,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.015599999576807022,"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/recurrent-neural-networks","display_name":"recurrent neural networks","score":0.8700000047683716},{"id":"https://openalex.org/keywords/long-context-modeling","display_name":"long-context modeling","score":0.8410000205039978},{"id":"https://openalex.org/keywords/state-credit","display_name":"state credit","score":0.621999979019165},{"id":"https://openalex.org/keywords/backpropagation-through-time","display_name":"backpropagation through time","score":0.8029999732971191},{"id":"https://openalex.org/keywords/length-extrapolation","display_name":"length extrapolation","score":0.625}],"concepts":[{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.6359000205993652},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5340999960899353},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.532800018787384},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4819999933242798},{"id":"https://openalex.org/C155032097","wikidata":"https://www.wikidata.org/wiki/Q798503","display_name":"Backpropagation","level":3,"score":0.4666000008583069},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.46299999952316284},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.43540000915527344},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.41429999470710754},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.4000000059604645},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.39329999685287476},{"id":"https://openalex.org/C40506919","wikidata":"https://www.wikidata.org/wiki/Q7452469","display_name":"Sequence learning","level":2,"score":0.3901999890804291},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.38370001316070557},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3792000114917755},{"id":"https://openalex.org/C191795146","wikidata":"https://www.wikidata.org/wiki/Q3878446","display_name":"Norm (philosophy)","level":2,"score":0.3314000070095062},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3310999870300293},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.33059999346733093},{"id":"https://openalex.org/C188116033","wikidata":"https://www.wikidata.org/wiki/Q2664563","display_name":"Q-learning","level":3,"score":0.3269999921321869},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.31839999556541443},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2935999929904938},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2847999930381775},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.2840999960899353},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.26919999718666077},{"id":"https://openalex.org/C104267543","wikidata":"https://www.wikidata.org/wiki/Q208163","display_name":"Signal processing","level":3,"score":0.26350000500679016},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.2619999945163727},{"id":"https://openalex.org/C38858127","wikidata":"https://www.wikidata.org/wiki/Q5441228","display_name":"Feed forward","level":2,"score":0.2533999979496002}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.09157","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.09157","pdf_url":"https://arxiv.org/pdf/2609.09157","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.2609.09157","doi":"https://doi.org/10.48550/arxiv.2609.09157","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.09157","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.09157","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.09157","pdf_url":"https://arxiv.org/pdf/2609.09157","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":[],"sustainable_development_goals_aurora":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7212071069.pdf","grobid_xml":"https://content.openalex.org/works/W7212071069.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recurrent":[0],"models":[1,10],"provide":[2],"a":[3,41],"natural":[4],"path":[5],"to":[6,75,101,131,149],"long-context":[7],"modeling,":[8],"yet":[9],"trained":[11],"with":[12,144],"backpropagation":[13],"through":[14,65,89],"time":[15],"(BPTT)":[16],"often":[17],"fail":[18],"beyond":[19,140],"their":[20],"training":[21,142,152],"horizon.":[22],"Classical":[23],"analyses":[24],"emphasize":[25],"gradients":[26],"that":[27,49],"vanish":[28],"or":[29],"explode":[30],"along":[31],"temporal":[32],"paths.":[33],"However,":[34],"dense":[35],"per-token":[36],"losses":[37,68],"can":[38],"still":[39],"train":[40],"shared":[42],"recurrent":[43,71],"rule":[44],"despite":[45],"severe":[46],"decay,":[47],"showing":[48],"decay":[50],"alone":[51],"does":[52],"not":[53],"determine":[54],"whether":[55],"learning":[56],"fails.":[57],"We":[58],"instead":[59],"study":[60],"state":[61,83],"credit:":[62],"the":[63,98,107,113,141,151],"signal":[64,100],"which":[66],"future":[67],"reach":[69],"earlier":[70],"states":[72],"before":[73],"contributing":[74],"parameter":[76],"updates.":[77],"Accordingly,":[78],"we":[79,128],"intervene":[80],"directly":[81],"on":[82],"credit":[84,126],"and":[85,121],"propose":[86],"Credit":[87],"Stabilization":[88],"Time":[90],"(CST).":[91],"During":[92],"backward":[93],"propagation,":[94],"CST":[95,130,137],"locally":[96],"rescales":[97],"state-credit":[99],"stabilize":[102],"its":[103],"norm":[104],"without":[105],"rotating":[106],"component":[108],"being":[109],"corrected,":[110],"while":[111],"leaving":[112],"forward":[114],"computation":[115],"unchanged.":[116],"Because":[117],"controlled":[118],"synthetic":[119],"tasks":[120],"real":[122],"data":[123],"exhibit":[124],"different":[125],"dynamics,":[127],"specialize":[129],"each":[132],"regime.":[133],"In":[134],"both":[135],"settings,":[136],"improves":[138],"performance":[139],"horizon,":[143],"gains":[145],"observed":[146],"at":[147],"up":[148],"128x":[150],"length.":[153]},"counts_by_year":[],"updated_date":"2026-09-30T09:27:27.532063","created_date":"2026-09-10T00:00:00"}