{"id":"https://openalex.org/W7213401090","doi":"https://doi.org/10.48550/arxiv.2609.16656","title":"Channel-Wise and Token-Aware Post-Training Quantization for Visual State Space Duality","display_name":"Channel-Wise and Token-Aware Post-Training Quantization for Visual State Space Duality","publication_year":2026,"publication_date":"2026-09-15","ids":{"openalex":"https://openalex.org/W7213401090","doi":"https://doi.org/10.48550/arxiv.2609.16656"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.16656","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.16656","pdf_url":"https://arxiv.org/pdf/2609.16656","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","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.16656","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5151908880","display_name":"Jonghyeon Lim","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I24062138","display_name":"Konkuk University","ror":"https://ror.org/025h1m602","country_code":"KR","type":"education","lineage":["https://openalex.org/I24062138"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Lim, Jonghyeon","raw_affiliation_strings":["Intelligent Image Processing Laboratory Department of Computer Science and Engineering , Konkuk University Seoul , Republic of Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Intelligent Image Processing Laboratory Department of Computer Science and Engineering , Konkuk University Seoul , Republic of Korea","institution_ids":["https://openalex.org/I24062138"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5024550177","display_name":"Changhoon Yim","orcid":"https://orcid.org/0000-0002-3826-6395","observed_orcids":["https://orcid.org/0000-0002-3826-6395"]},"institutions":[{"id":"https://openalex.org/I24062138","display_name":"Konkuk University","ror":"https://ror.org/025h1m602","country_code":"KR","type":"education","lineage":["https://openalex.org/I24062138"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Yim, Changhoon","raw_affiliation_strings":["Intelligent Image Processing Laboratory Department of Computer Science and Engineering , Konkuk University Seoul , Republic of Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Intelligent Image Processing Laboratory Department of Computer Science and Engineering , Konkuk University Seoul , Republic of Korea","institution_ids":["https://openalex.org/I24062138"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I24062138"],"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9994999766349792,"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":0.9994999766349792,"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":0.00019999999494757503,"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/T11605","display_name":"Visual Attention and Saliency Detection","score":0.0,"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/post-training-quantization","display_name":"post-training quantization","score":0.9679999947547913},{"id":"https://openalex.org/keywords/visual-state-space-model","display_name":"visual state space model","score":0.7760000228881836},{"id":"https://openalex.org/keywords/activation-quantization","display_name":"activation quantization","score":0.7149999737739563},{"id":"https://openalex.org/keywords/mamba","display_name":"Mamba","score":0.5820000171661377},{"id":"https://openalex.org/keywords/low-bit-precision","display_name":"low-bit precision","score":0.5350000262260437},{"id":"https://openalex.org/keywords/object-detection","display_name":"object detection","score":0.4970000088214874},{"id":"https://openalex.org/keywords/instance-segmentation","display_name":"instance segmentation","score":0.8740000128746033},{"id":"https://openalex.org/keywords/semantic-segmentation","display_name":"semantic segmentation","score":0.9950000047683716}],"concepts":[{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.7652000188827515},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5527999997138977},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5115000009536743},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.4959000051021576},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.40860000252723694},{"id":"https://openalex.org/C49766605","wikidata":"https://www.wikidata.org/wiki/Q207643","display_name":"Linear map","level":2,"score":0.3930000066757202},{"id":"https://openalex.org/C89600930","wikidata":"https://www.wikidata.org/wiki/Q1423946","display_name":"Segmentation","level":2,"score":0.3865000009536743},{"id":"https://openalex.org/C2778023678","wikidata":"https://www.wikidata.org/wiki/Q554403","display_name":"Duality (order theory)","level":2,"score":0.36660000681877136},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3653999865055084},{"id":"https://openalex.org/C2776848632","wikidata":"https://www.wikidata.org/wiki/Q853463","display_name":"Clipping (morphology)","level":2,"score":0.3366999924182892},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3361000120639801},{"id":"https://openalex.org/C72434380","wikidata":"https://www.wikidata.org/wiki/Q230930","display_name":"State space","level":2,"score":0.33000001311302185},{"id":"https://openalex.org/C207363949","wikidata":"https://www.wikidata.org/wiki/Q462915","display_name":"Visual space","level":3,"score":0.32850000262260437},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.3231000006198883},{"id":"https://openalex.org/C62799726","wikidata":"https://www.wikidata.org/wiki/Q190056","display_name":"Hilbert space","level":2,"score":0.32100000977516174},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.30379998683929443},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2971000075340271},{"id":"https://openalex.org/C2781238097","wikidata":"https://www.wikidata.org/wiki/Q175026","display_name":"Object (grammar)","level":2,"score":0.2741999924182892},{"id":"https://openalex.org/C112680207","wikidata":"https://www.wikidata.org/wiki/Q714886","display_name":"Regular polygon","level":2,"score":0.2614000141620636}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.16656","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.16656","pdf_url":"https://arxiv.org/pdf/2609.16656","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.2609.16656","doi":"https://doi.org/10.48550/arxiv.2609.16656","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.16656","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:2609.16656","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.16656","pdf_url":"https://arxiv.org/pdf/2609.16656","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":[{"id":"https://openalex.org/F4320320671","display_name":"National Research Foundation","ror":"https://ror.org/05s0g1g46"},{"id":"https://openalex.org/F4320322120","display_name":"National Research Foundation of Korea","ror":"https://ror.org/013aysd81"},{"id":"https://openalex.org/F4320328359","display_name":"Ministry of Science and ICT, South Korea","ror":"https://ror.org/01wpjm123"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7213401090.pdf","grobid_xml":"https://content.openalex.org/works/W7213401090.grobid-xml"},"referenced_works_count":3,"referenced_works":["https://openalex.org/W2242818861","https://openalex.org/W4403622927","https://openalex.org/W4406793343"],"related_works":["https://openalex.org/W4406793343","https://openalex.org/W4415524158","https://openalex.org/W7160696699","https://openalex.org/W7155364465","https://openalex.org/W4375868854","https://openalex.org/W7133220258","https://openalex.org/W4400575198","https://openalex.org/W4388843165","https://openalex.org/W7133229067","https://openalex.org/W6948382120"],"abstract_inverted_index":{"State":[0,25],"space":[1],"models":[2],"(SSMs),":[3],"particularly":[4],"Mamba,":[5],"have":[6,15],"emerged":[7],"as":[8,49],"efficient":[9],"alternatives":[10],"to":[11,18,57,144,169],"attention-based":[12],"architectures":[13],"and":[14,23,67,100,116,125,149,157],"been":[16],"extended":[17],"vision":[19],"through":[20],"ViM,":[21],"VMamba,":[22],"Visual":[24],"Space":[26],"Duality":[27],"(VSSD).":[28],"Yet":[29],"the":[30,50,72,91,96,118,131,140],"low-bit":[31,52],"post-training":[32],"quantization":[33,48,142],"(PTQ)":[34],"behavior":[35],"of":[36],"VSSD":[37,145],"remains":[38,126],"insufficiently":[39],"understood.":[40],"A":[41],"weight-activation":[42],"split":[43],"on":[44,90,147],"VSSD-Tiny":[45],"identifies":[46],"activation":[47],"dominant":[51],"bottleneck,":[53],"while":[54],"representative":[55],"inputs":[56],"selected":[58,97],"VSSD-backbone":[59],"linear":[60,93,98],"layers":[61,99],"exhibit":[62],"strong":[63,152],"channel-wise":[64],"magnitude":[65],"variation":[66],"token-localized":[68],"extremes.":[69],"We":[70],"propose":[71],"Channel-wise":[73],"Token-balanced":[74],"Output-Aware":[75],"Clipping":[76],"(CTOAC)":[77],"method,":[78],"which":[79],"learns":[80],"per-input-channel":[81],"clipping":[82],"bounds":[83],"by":[84],"minimizing":[85],"a":[86],"token-balanced":[87],"reconstruction":[88],"loss":[89],"corresponding":[92],"outputs.":[94],"Only":[95],"their":[101,110],"input":[102],"activations":[103],"are":[104],"quantized;":[105],"other":[106],"backbone":[107],"operations":[108],"retain":[109],"original":[111],"precision.":[112],"Across":[113],"VSSD-Tiny,":[114],"VSSD-Small,":[115],"VSSD-Base,":[117],"proposed":[119],"CTOAC":[120],"method":[121],"retains":[122],"ImageNet-1K":[123],"accuracy":[124],"substantially":[127],"more":[128,135],"robust":[129],"than":[130],"evaluated":[132],"baselines":[133],"at":[134],"aggressive":[136],"precision":[137],"settings.":[138],"Applying":[139],"same":[141],"scope":[143],"backbones":[146],"COCO":[148],"ADE20K":[150],"preserves":[151],"object":[153],"detection,":[154],"instance":[155],"segmentation,":[156],"semantic":[158],"segmentation":[159],"performance.":[160],"An":[161],"optimized":[162],"RTX":[163],"4090":[164],"deployment":[165],"configuration":[166],"achieves":[167],"up":[168],"1.42x":[170],"end-to-end":[171],"speedup":[172],"over":[173],"FP32.":[174]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-17T00:00:00"}