{"id":"https://openalex.org/W3181925591","doi":"https://doi.org/10.48550/arxiv.2107.04735","title":"Local-to-Global Self-Attention in Vision Transformers","display_name":"Local-to-Global Self-Attention in Vision Transformers","publication_year":2021,"publication_date":"2021-07-10","ids":{"openalex":"https://openalex.org/W3181925591","doi":"https://doi.org/10.48550/arxiv.2107.04735","mag":"3181925591"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2107.04735","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2107.04735","pdf_url":"https://arxiv.org/pdf/2107.04735","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/2107.04735","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100782322","display_name":"Jinpeng Li","orcid":"https://orcid.org/0000-0003-1752-8883","observed_orcids":["https://orcid.org/0000-0003-1752-8883","https://orcid.org/0000-0002-1538-7274"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Jinpeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048907728","display_name":"Yichao Yan","orcid":"https://orcid.org/0000-0003-3209-8965","observed_orcids":["https://orcid.org/0000-0003-3209-8965"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yan, Yichao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019092879","display_name":"Shengcai Liao","orcid":"https://orcid.org/0000-0001-8941-2295","observed_orcids":["https://orcid.org/0000-0001-8941-2295"]},"institutions":[{"id":"https://openalex.org/I3132238960","display_name":"Institute of Electrical and Electronics Engineers","ror":"https://ror.org/01n002310","country_code":"US","type":"education","lineage":["https://openalex.org/I3132238960"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Liao, Shengcai","raw_affiliation_strings":["Institute of Electrical and Electronics Engineers"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Institute of Electrical and Electronics Engineers","institution_ids":["https://openalex.org/I3132238960"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019708391","display_name":"Xiaokang Yang","orcid":"https://orcid.org/0000-0003-4029-3322","observed_orcids":["https://orcid.org/0000-0003-4029-3322"]},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang, Xiaokang","raw_affiliation_strings":["Shanghai Jiao Tong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5082634513","display_name":"Ling Shao","orcid":"https://orcid.org/0000-0002-8264-6117","observed_orcids":["https://orcid.org/0000-0002-8264-6117"]},"institutions":[{"id":"https://openalex.org/I3132238960","display_name":"Institute of Electrical and Electronics Engineers","ror":"https://ror.org/01n002310","country_code":"US","type":"education","lineage":["https://openalex.org/I3132238960"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Shao, Ling","raw_affiliation_strings":["Institute of Electrical and Electronics Engineers"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Institute of Electrical and Electronics Engineers","institution_ids":["https://openalex.org/I3132238960"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":25,"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":0.98580002784729,"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":0.98580002784729,"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/T11605","display_name":"Visual Attention and Saliency Detection","score":0.0071000000461936,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.0031999999191612005,"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/vision-transformer","display_name":"vision transformer","score":0.9210000038146973},{"id":"https://openalex.org/keywords/self-attention","display_name":"self-attention","score":0.8669999837875366},{"id":"https://openalex.org/keywords/hierarchical-attention","display_name":"hierarchical attention","score":0.6019999980926514},{"id":"https://openalex.org/keywords/computational-efficiency","display_name":"computational efficiency","score":0.6800000071525574},{"id":"https://openalex.org/keywords/image-classification","display_name":"image classification","score":0.9539999961853027},{"id":"https://openalex.org/keywords/semantic-segmentation","display_name":"semantic segmentation","score":0.7940000295639038}],"concepts":[{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.4180598855018616},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.33642834424972534},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.15173554420471191},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.12154600024223328},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0555344820022583}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2107.04735","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2107.04735","pdf_url":"https://arxiv.org/pdf/2107.04735","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.2107.04735","doi":"https://doi.org/10.48550/arxiv.2107.04735","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2107.04735","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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2107.04735","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2107.04735","pdf_url":"https://arxiv.org/pdf/2107.04735","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":[{"score":0.41999998688697815,"id":"https://metadata.un.org/sdg/9","display_name":"Industry, innovation and infrastructure"}],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":57,"referenced_works":["https://openalex.org/W1686810756","https://openalex.org/W1861492603","https://openalex.org/W2097117768","https://openalex.org/W2108598243","https://openalex.org/W2125556102","https://openalex.org/W2151103935","https://openalex.org/W2161969291","https://openalex.org/W2168356304","https://openalex.org/W2183341477","https://openalex.org/W2194775991","https://openalex.org/W2307770531","https://openalex.org/W2412782625","https://openalex.org/W2418011751","https://openalex.org/W2490270993","https://openalex.org/W2507296351","https://openalex.org/W2560023338","https://openalex.org/W2565639579","https://openalex.org/W2612445135","https://openalex.org/W2626778328","https://openalex.org/W2736981815","https://openalex.org/W2884822772","https://openalex.org/W2895126795","https://openalex.org/W2916798096","https://openalex.org/W2936599103","https://openalex.org/W2937843571","https://openalex.org/W2946948417","https://openalex.org/W2955058313","https://openalex.org/W2963163009","https://openalex.org/W2963341956","https://openalex.org/W2963446712","https://openalex.org/W2964309882","https://openalex.org/W2964350391","https://openalex.org/W2965373594","https://openalex.org/W2970419734","https://openalex.org/W2976122994","https://openalex.org/W2981851019","https://openalex.org/W2991062542","https://openalex.org/W3016265891","https://openalex.org/W3030520226","https://openalex.org/W3033210410","https://openalex.org/W3035022492","https://openalex.org/W3092462694","https://openalex.org/W3098903812","https://openalex.org/W3107036272","https://openalex.org/W3107634219","https://openalex.org/W3116489684","https://openalex.org/W3119786062","https://openalex.org/W3121523901","https://openalex.org/W3131500599","https://openalex.org/W3138516171","https://openalex.org/W3139049060","https://openalex.org/W3151130473","https://openalex.org/W3152694111","https://openalex.org/W3170841864","https://openalex.org/W3171125843","https://openalex.org/W3204076343","https://openalex.org/W4389666313"],"related_works":["https://openalex.org/W4361010055","https://openalex.org/W4225677168","https://openalex.org/W3176153963","https://openalex.org/W4387883160","https://openalex.org/W3170642968","https://openalex.org/W4385195799","https://openalex.org/W3166942762","https://openalex.org/W4366352791","https://openalex.org/W4287124731","https://openalex.org/W4411275199"],"abstract_inverted_index":{"Transformers":[0],"have":[1],"demonstrated":[2],"great":[3],"potential":[4],"in":[5,15,46,67,83,91],"computer":[6],"vision":[7],"tasks.":[8],"To":[9],"avoid":[10],"dense":[11],"computations":[12],"of":[13,57],"self-attentions":[14,28],"high-resolution":[16],"visual":[17],"data,":[18],"some":[19],"recent":[20],"Transformer":[21],"models":[22],"adopt":[23],"a":[24,54,80],"hierarchical":[25],"design,":[26],"where":[27],"are":[29],"only":[30],"computed":[31],"within":[32],"local":[33],"windows.":[34],"This":[35],"design":[36,53],"significantly":[37],"improves":[38],"the":[39,58],"efficiency":[40],"but":[41],"lacks":[42],"global":[43],"feature":[44],"reasoning":[45,63],"early":[47],"stages.":[48],"In":[49],"this":[50],"work,":[51],"we":[52],"multi-path":[55],"structure":[56],"Transformer,":[59],"which":[60],"enables":[61],"local-to-global":[62],"at":[64,101],"multiple":[65],"granularities":[66],"each":[68],"stage.":[69],"The":[70],"proposed":[71],"framework":[72],"is":[73,99],"computationally":[74],"efficient":[75],"and":[76,95],"highly":[77],"effective.":[78],"With":[79],"marginal":[81],"increasement":[82],"computational":[84],"overhead,":[85],"our":[86],"model":[87],"achieves":[88],"notable":[89],"improvements":[90],"both":[92],"image":[93],"classification":[94],"semantic":[96],"segmentation.":[97],"Code":[98],"available":[100],"https://github.com/ljpadam/LG-Transformer":[102]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":8},{"year":2023,"cited_by_count":8},{"year":2022,"cited_by_count":6},{"year":2021,"cited_by_count":1}],"updated_date":"2026-10-16T09:59:22.235173","created_date":"2021-07-19T00:00:00"}