{"id":"https://openalex.org/W7128775486","doi":"https://doi.org/10.48550/arxiv.2602.11698","title":"SpiralFormer: Looped Transformers Can Learn Hierarchical Dependencies via Multi-Resolution Recursion","display_name":"SpiralFormer: Looped Transformers Can Learn Hierarchical Dependencies via Multi-Resolution Recursion","publication_year":2026,"publication_date":"2026-02-12","ids":{"openalex":"https://openalex.org/W7128775486","doi":"https://doi.org/10.48550/arxiv.2602.11698"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2602.11698","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.11698","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2602.11698","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5014148417","display_name":"Chengting Yu","orcid":"https://orcid.org/0009-0007-7210-879X"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Yu, Chengting","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125934430","display_name":"Xiaobo Shu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shu, Xiaobo","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5038360189","display_name":"Yadao Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yadao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121579926","display_name":"Yizhen Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yizhen","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125947395","display_name":"Haoyi Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Haoyi","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125952622","display_name":"You Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, You","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023108787","display_name":"Rujiao Long","orcid":"https://orcid.org/0000-0003-1330-3193"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Long, Rujiao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125967635","display_name":"Ziheng Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Ziheng","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125892916","display_name":"Yuchi Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Yuchi","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124038958","display_name":"Wenbo Su","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Su, Wenbo","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5125961400","display_name":"Bo Zheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Bo","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":11,"corresponding_author_ids":["https://openalex.org/A5014148417"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.2590000033378601,"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"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.2590000033378601,"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"}},{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.16760000586509705,"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"}},{"id":"https://openalex.org/T10719","display_name":"3D Shape Modeling and Analysis","score":0.06589999794960022,"subfield":{"id":"https://openalex.org/subfields/2206","display_name":"Computational Mechanics"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/recursion","display_name":"Recursion (computer science)","score":0.8575999736785889},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.6744999885559082},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.6031000018119812},{"id":"https://openalex.org/keywords/divide-and-conquer-algorithms","display_name":"Divide and conquer algorithms","score":0.38089999556541443},{"id":"https://openalex.org/keywords/iterative-method","display_name":"Iterative method","score":0.3190999925136566},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.30979999899864197}],"concepts":[{"id":"https://openalex.org/C168773036","wikidata":"https://www.wikidata.org/wiki/Q264164","display_name":"Recursion (computer science)","level":2,"score":0.8575999736785889},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.6744999885559082},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6335999965667725},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.6031000018119812},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5220999717712402},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.5067999958992004},{"id":"https://openalex.org/C71559656","wikidata":"https://www.wikidata.org/wiki/Q671298","display_name":"Divide and conquer algorithms","level":2,"score":0.38089999556541443},{"id":"https://openalex.org/C159694833","wikidata":"https://www.wikidata.org/wiki/Q2321565","display_name":"Iterative method","level":2,"score":0.3190999925136566},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.30979999899864197},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.30300000309944153},{"id":"https://openalex.org/C3020199158","wikidata":"https://www.wikidata.org/wiki/Q210521","display_name":"High resolution","level":2,"score":0.2809000015258789},{"id":"https://openalex.org/C2779982483","wikidata":"https://www.wikidata.org/wiki/Q6094420","display_name":"Iterative refinement","level":2,"score":0.27480000257492065},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.26019999384880066}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2602.11698","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.11698","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2602.11698","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.11698","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recursive":[0],"(looped)":[1],"Transformers":[2,28],"decouple":[3],"computational":[4],"depth":[5,8],"from":[6,124],"parameter":[7,111],"by":[9,99],"repeatedly":[10],"applying":[11],"shared":[12],"layers,":[13],"providing":[14],"an":[15],"explicit":[16],"architectural":[17],"primitive":[18],"for":[19,135],"iterative":[20],"refinement":[21],"and":[22,112,118],"latent":[23,66],"reasoning.":[24],"However,":[25],"early":[26],"looped":[27,75,117],"often":[29],"underperform":[30],"non-recursive":[31],"baselines":[32,120],"of":[33,62],"equal":[34],"compute.":[35],"While":[36],"recent":[37],"literature":[38],"has":[39],"introduced":[40],"more":[41],"effective":[42],"recursion":[43,83,91],"mechanisms":[44],"to":[45,95,126],"mitigate":[46],"this":[47,69],"gap,":[48],"existing":[49],"architectures":[50],"still":[51],"operate":[52],"at":[53],"a":[54,74,81,132],"fixed,":[55],"full-token":[56],"resolution,":[57],"neglecting":[58],"the":[59,93],"potential":[60,133],"efficiency":[61,114],"computing":[63],"over":[64],"compressed":[65],"representations.":[67],"In":[68],"paper,":[70],"we":[71],"propose":[72],"SpiralFormer,":[73],"Transformer":[76],"that":[77,89],"executes":[78],"recurrence":[79],"under":[80],"multi-resolution":[82,90],"schedule.":[84],"We":[85],"provide":[86],"probing":[87],"evidence":[88],"enables":[92],"model":[94,122],"learn":[96],"hierarchical":[97],"dependencies":[98],"inducing":[100],"iteration-wise":[101],"functional":[102],"specialization":[103],"across":[104,121],"different":[105],"scales.":[106],"Empirically,":[107],"SpiralFormer":[108],"achieves":[109],"better":[110],"compute":[113],"than":[115],"both":[116],"non-looped":[119],"scales":[123],"160M":[125],"1.4B,":[127],"establishing":[128],"sequence":[129],"resolution":[130],"as":[131],"axis":[134],"scaling":[136],"recursive":[137],"architectures.":[138]},"counts_by_year":[],"updated_date":"2026-02-14T06:27:19.455381","created_date":"2026-02-14T00:00:00"}