{"id":"https://openalex.org/W7224749111","doi":"https://doi.org/10.48550/arxiv.2610.40647","title":"Read, Allocate, Route: Role-Coupled Tokenization for Pixel-Space Generative Transformers","display_name":"Read, Allocate, Route: Role-Coupled Tokenization for Pixel-Space Generative Transformers","publication_year":2026,"publication_date":"2026-10-29","ids":{"openalex":"https://openalex.org/W7224749111","doi":"https://doi.org/10.48550/arxiv.2610.40647"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.40647","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.40647","pdf_url":"https://arxiv.org/pdf/2610.40647","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/2610.40647","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137919345","display_name":"Jiongfan Yang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang, Jiongfan","raw_affiliation_strings":["Peking University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Peking University ,","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5027750483","display_name":"Weichen Fan","orcid":"https://orcid.org/0000-0003-0041-955X","observed_orcids":["https://orcid.org/0000-0003-0041-955X"]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]},{"id":"https://openalex.org/I4401727009","display_name":"Sensetime (China)","ror":"https://ror.org/049wsmj07","country_code":"CN","type":"company","lineage":["https://openalex.org/I4401727009"]}],"countries":["CN","SG"],"is_corresponding":false,"raw_author_name":"Fan, Weichen","raw_affiliation_strings":["Nanyang Technological University , 3 SenseTime ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanyang Technological University , 3 SenseTime ,","institution_ids":["https://openalex.org/I172675005","https://openalex.org/I4401727009"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5015049935","display_name":"Haiwen Diao","orcid":"https://orcid.org/0000-0002-4156-5417","observed_orcids":["https://orcid.org/0000-0002-4156-5417"]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]},{"id":"https://openalex.org/I4401727009","display_name":"Sensetime (China)","ror":"https://ror.org/049wsmj07","country_code":"CN","type":"company","lineage":["https://openalex.org/I4401727009"]}],"countries":["CN","SG"],"is_corresponding":false,"raw_author_name":"Diao, Haiwen","raw_affiliation_strings":["Nanyang Technological University , 3 SenseTime ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanyang Technological University , 3 SenseTime ,","institution_ids":["https://openalex.org/I172675005","https://openalex.org/I4401727009"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054947207","display_name":"Zhongang Cai","orcid":"https://orcid.org/0000-0002-1810-3855","observed_orcids":["https://orcid.org/0000-0002-1810-3855"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cai, Zhongang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5160214028","display_name":"Ziwei Liu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]},{"id":"https://openalex.org/I4401727009","display_name":"Sensetime (China)","ror":"https://ror.org/049wsmj07","country_code":"CN","type":"company","lineage":["https://openalex.org/I4401727009"]}],"countries":["CN","SG"],"is_corresponding":false,"raw_author_name":"Liu, Ziwei","raw_affiliation_strings":["Nanyang Technological University , 3 SenseTime ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanyang Technological University , 3 SenseTime ,","institution_ids":["https://openalex.org/I172675005","https://openalex.org/I4401727009"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.9970999956130981,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.9970999956130981,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.0019000000320374966,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.00019999999494757503,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/international-attention","display_name":"global attention","score":0.4620000123977661},{"id":"https://openalex.org/keywords/image-generation","display_name":"image generation","score":0.47200000286102295},{"id":"https://openalex.org/keywords/text-to-image-generation","display_name":"text-to-image generation","score":0.9359999895095825},{"id":"https://openalex.org/keywords/computational-efficiency","display_name":"computational efficiency","score":0.5659999847412109}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6721000075340271},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.5874999761581421},{"id":"https://openalex.org/C45340560","wikidata":"https://www.wikidata.org/wiki/Q215382","display_name":"Disjoint sets","level":2,"score":0.5133000016212463},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.48399999737739563},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4544000029563904},{"id":"https://openalex.org/C160633673","wikidata":"https://www.wikidata.org/wiki/Q355198","display_name":"Pixel","level":2,"score":0.4523000121116638},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.4113999903202057},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3529999852180481},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3522999882698059},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.33239999413490295},{"id":"https://openalex.org/C97256817","wikidata":"https://www.wikidata.org/wiki/Q1462316","display_name":"Spurious relationship","level":2,"score":0.32580000162124634},{"id":"https://openalex.org/C19743564","wikidata":"https://www.wikidata.org/wiki/Q25378119","display_name":"Flicker","level":2,"score":0.31630000472068787},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.2964000105857849},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.26179999113082886},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.25859999656677246}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.40647","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.40647","pdf_url":"https://arxiv.org/pdf/2610.40647","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.2610.40647","doi":"https://doi.org/10.48550/arxiv.2610.40647","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.40647","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":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2610.40647","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.40647","pdf_url":"https://arxiv.org/pdf/2610.40647","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":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7224749111.pdf","grobid_xml":"https://content.openalex.org/works/W7224749111.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7170095198","https://openalex.org/W7212704818","https://openalex.org/W7169834629","https://openalex.org/W7223781274","https://openalex.org/W7214920779","https://openalex.org/W4405355469","https://openalex.org/W4416686928","https://openalex.org/W7201842879","https://openalex.org/W7106789420","https://openalex.org/W7130710665"],"abstract_inverted_index":{"Pixel-space":[0],"generative":[1,137],"Transformers":[2],"typically":[3],"use":[4],"one":[5],"token":[6],"per":[7],"image":[8],"patch,":[9],"so":[10],"capturing":[11],"finer":[12,26],"detail":[13,27,43,62,117],"requires":[14],"more":[15,32],"tokens":[16],"and":[17,140,176],"makes":[18],"global":[19,33,39,68,103,121],"attention":[20,69],"increasingly":[21],"expensive.":[22],"We":[23],"observe":[24],"that":[25,59],"does":[28],"not":[29],"necessarily":[30],"require":[31],"communication:":[34],"low-frequency":[35],"structure":[36],"benefits":[37],"from":[38,63,91,98],"exchange,":[40],"while":[41,105,135,159],"high-frequency":[42],"can":[44],"be":[45],"modeled":[46],"locally.":[47,109],"Based":[48],"on":[49,194],"this":[50,87],"insight,":[51],"we":[52],"introduce":[53],"SpecSlot,":[54],"a":[55,81,128,150,155],"pixel-space":[56],"rectified-flow":[57],"Transformer":[58],"decouples":[60],"output":[61],"the":[64,67,95,112,120,132,136,170],"length":[65],"of":[66,84,182],"sequence.":[70,122],"SpecSlot":[71,148,191],"expands":[72],"each":[73,79,99],"spatial":[74,124],"anchor":[75,89,100],"into":[76],"multiple":[77],"slots,":[78],"assigned":[80],"disjoint":[82],"band":[83],"DCT":[85],"coefficients;":[86],"separates":[88],"density":[90],"spectral":[92,116,133],"subdivision.":[93],"Only":[94],"lowest-frequency":[96],"slot":[97],"participates":[101],"in":[102,143],"attention,":[104],"all":[106],"slots":[107],"interact":[108],"This":[110],"allows":[111],"model":[113],"to":[114,131,154],"increase":[115],"without":[118],"enlarging":[119],"A":[123],"head":[125],"further":[126],"adds":[127],"local":[129],"correction":[130],"prediction,":[134],"state,":[138],"target,":[139],"loss":[141],"remain":[142],"pixel":[144],"space.":[145],"On":[146,162,188],"ImageNet-256,":[147],"achieves":[149],"comparable":[151],"final":[152],"FID":[153,166,181],"FLOPs-matched":[156],"JiT":[157],"control":[158,171],"converging":[160],"faster.":[161],"ImageNet-512,":[163],"it":[164],"improves":[165],"by":[167,196],"2.84":[168],"over":[169],"with":[172],"3.9%":[173],"fewer":[174],"FLOPs,":[175],"comes":[177],"within":[178],"only":[179],"0.69":[180],"JiT-B/16":[183],"at":[184,198],"38.2%":[185],"lower":[186,200],"cost.":[187,202],"text-to-image":[189],"generation,":[190],"surpasses":[192],"JiT-B/8":[193],"DPG-Bench":[195],"1.06":[197],"40.6%":[199],"forward":[201]},"counts_by_year":[],"updated_date":"2026-11-04T10:54:05.112534","created_date":"2026-10-31T00:00:00"}