{"id":"https://openalex.org/W7214736285","doi":"https://doi.org/10.48550/arxiv.2609.33269","title":"Q-WAM: 4-Bit Quantization of World Action Models with Action-Subspace Protection","display_name":"Q-WAM: 4-Bit Quantization of World Action Models with Action-Subspace Protection","publication_year":2026,"publication_date":"2026-09-27","ids":{"openalex":"https://openalex.org/W7214736285","doi":"https://doi.org/10.48550/arxiv.2609.33269"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.33269","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.33269","pdf_url":"https://arxiv.org/pdf/2609.33269","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":"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/2609.33269","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135994307","display_name":"Arash Akbari","orcid":"https://orcid.org/0009-0006-1846-0672","observed_orcids":["https://orcid.org/0009-0006-1846-0672"]},"institutions":[{"id":"https://openalex.org/I12912129","display_name":"Northeastern University","ror":"https://ror.org/04t5xt781","country_code":"US","type":"education","lineage":["https://openalex.org/I12912129"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Akbari, Arash","raw_affiliation_strings":["Northeastern University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Northeastern University","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135957646","display_name":"Arman Akbari","orcid":"https://orcid.org/0009-0009-6835-030X","observed_orcids":["https://orcid.org/0009-0009-6835-030X"]},"institutions":[{"id":"https://openalex.org/I12912129","display_name":"Northeastern University","ror":"https://ror.org/04t5xt781","country_code":"US","type":"education","lineage":["https://openalex.org/I12912129"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Akbari, Arman","raw_affiliation_strings":["Northeastern University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Northeastern University","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125587858","display_name":"Jingwu Luo","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I12912129","display_name":"Northeastern University","ror":"https://ror.org/04t5xt781","country_code":"US","type":"education","lineage":["https://openalex.org/I12912129"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Luo, Jingwu","raw_affiliation_strings":["Northeastern University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Northeastern University","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153483684","display_name":"Yuhao Lei","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I12912129","display_name":"Northeastern University","ror":"https://ror.org/04t5xt781","country_code":"US","type":"education","lineage":["https://openalex.org/I12912129"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Lei, Yuhao","raw_affiliation_strings":["Northeastern University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Northeastern University","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153455341","display_name":"Yi Gao","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I12912129","display_name":"Northeastern University","ror":"https://ror.org/04t5xt781","country_code":"US","type":"education","lineage":["https://openalex.org/I12912129"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Gao, Yi","raw_affiliation_strings":["Northeastern University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Northeastern University","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153518928","display_name":"Weiwei Chen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210109081","display_name":"Embody (France)","ror":"https://ror.org/01sy9dm60","country_code":"FR","type":"company","lineage":["https://openalex.org/I4210109081"]}],"countries":["FR"],"is_corresponding":false,"raw_author_name":"Chen, Weiwei","raw_affiliation_strings":["EmbodyX"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"EmbodyX","institution_ids":["https://openalex.org/I4210109081"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153429050","display_name":"Xuan Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I12912129","display_name":"Northeastern University","ror":"https://ror.org/04t5xt781","country_code":"US","type":"education","lineage":["https://openalex.org/I12912129"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhang, Xuan","raw_affiliation_strings":["Northeastern University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Northeastern University","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5065889904","display_name":"Zhenman Fang","orcid":"https://orcid.org/0000-0003-0603-9697","observed_orcids":["https://orcid.org/0000-0003-0603-9697"]},"institutions":[{"id":"https://openalex.org/I130238516","display_name":"University of Minnesota","ror":"https://ror.org/017zqws13","country_code":"US","type":"education","lineage":["https://openalex.org/I130238516"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Fang, Zhenman","raw_affiliation_strings":["University of Minnesota -Twin Cities"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Minnesota -Twin Cities","institution_ids":["https://openalex.org/I130238516"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5007300551","display_name":"Geng Yuan","orcid":"https://orcid.org/0000-0001-9844-992X","observed_orcids":["https://orcid.org/0000-0001-9844-992X"]},"institutions":[{"id":"https://openalex.org/I165733156","display_name":"University of Georgia","ror":"https://ror.org/00te3t702","country_code":"US","type":"education","lineage":["https://openalex.org/I165733156"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yuan, Geng","raw_affiliation_strings":["University of Georgia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Georgia","institution_ids":["https://openalex.org/I165733156"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100651384","display_name":"Yanzhi Wang","orcid":"https://orcid.org/0000-0002-3024-7990","observed_orcids":["https://orcid.org/0000-0002-3024-7990","https://orcid.org/0009-0009-3924-6466","https://orcid.org/0009-0008-4749-5860"]},"institutions":[{"id":"https://openalex.org/I12912129","display_name":"Northeastern University","ror":"https://ror.org/04t5xt781","country_code":"US","type":"education","lineage":["https://openalex.org/I12912129"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Wang, Yanzhi","raw_affiliation_strings":["Northeastern University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Northeastern University","institution_ids":["https://openalex.org/I12912129"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9624000191688538,"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":0.9624000191688538,"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/T10653","display_name":"Robot Manipulation and Learning","score":0.0071000000461936,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.002400000113993883,"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/world-action-models","display_name":"world action models","score":0.8949999809265137},{"id":"https://openalex.org/keywords/post-training-quantization","display_name":"post-training quantization","score":0.8870000243186951},{"id":"https://openalex.org/keywords/action-generation","display_name":"action generation","score":0.44200000166893005},{"id":"https://openalex.org/keywords/robotic-manipulation","display_name":"robotic manipulation","score":0.6940000057220459},{"id":"https://openalex.org/keywords/low-rank-approximation","display_name":"low-rank approximation","score":0.7120000123977661},{"id":"https://openalex.org/keywords/expert-selection","display_name":"expert selection","score":0.48399999737739563}],"concepts":[{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.6629999876022339},{"id":"https://openalex.org/C36299963","wikidata":"https://www.wikidata.org/wiki/Q1369844","display_name":"Observability","level":2,"score":0.5472000241279602},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5340999960899353},{"id":"https://openalex.org/C3770464","wikidata":"https://www.wikidata.org/wiki/Q775963","display_name":"Smoothing","level":2,"score":0.52920001745224},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4327000081539154},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.41920000314712524},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.41909998655319214},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3898000121116638},{"id":"https://openalex.org/C136625980","wikidata":"https://www.wikidata.org/wiki/Q663208","display_name":"Rounding","level":2,"score":0.36730000376701355},{"id":"https://openalex.org/C90509273","wikidata":"https://www.wikidata.org/wiki/Q11012","display_name":"Robot","level":2,"score":0.3662000000476837},{"id":"https://openalex.org/C60692881","wikidata":"https://www.wikidata.org/wiki/Q584529","display_name":"Humanoid robot","level":3,"score":0.33869999647140503},{"id":"https://openalex.org/C153207627","wikidata":"https://www.wikidata.org/wiki/Q863873","display_name":"Code word","level":3,"score":0.3285999894142151},{"id":"https://openalex.org/C79337645","wikidata":"https://www.wikidata.org/wiki/Q779824","display_name":"Outlier","level":2,"score":0.326200008392334},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3206999897956848},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.2793000042438507},{"id":"https://openalex.org/C128107574","wikidata":"https://www.wikidata.org/wiki/Q182003","display_name":"Injective function","level":2,"score":0.27300000190734863},{"id":"https://openalex.org/C159694833","wikidata":"https://www.wikidata.org/wiki/Q2321565","display_name":"Iterative method","level":2,"score":0.26739999651908875},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.2669000029563904},{"id":"https://openalex.org/C176809094","wikidata":"https://www.wikidata.org/wiki/Q15401496","display_name":"Traverse","level":2,"score":0.26030001044273376},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.2513999938964844}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.33269","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.33269","pdf_url":"https://arxiv.org/pdf/2609.33269","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.33269","doi":"https://doi.org/10.48550/arxiv.2609.33269","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.33269","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:2609.33269","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.33269","pdf_url":"https://arxiv.org/pdf/2609.33269","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214736285.pdf","grobid_xml":"https://content.openalex.org/works/W7214736285.grobid-xml"},"referenced_works_count":26,"referenced_works":["https://openalex.org/W3122499249","https://openalex.org/W4303647933","https://openalex.org/W4307934016","https://openalex.org/W4309591680","https://openalex.org/W4379260375","https://openalex.org/W4390874074","https://openalex.org/W4399115604","https://openalex.org/W4404363168","https://openalex.org/W4404388936","https://openalex.org/W4415183980","https://openalex.org/W4415796210","https://openalex.org/W4415797048","https://openalex.org/W4417453644","https://openalex.org/W6929480010","https://openalex.org/W7125496427","https://openalex.org/W7126180855","https://openalex.org/W7127648241","https://openalex.org/W7130552073","https://openalex.org/W7131410930","https://openalex.org/W7138923929","https://openalex.org/W7139013557","https://openalex.org/W7162458638","https://openalex.org/W7163686620","https://openalex.org/W7165373886","https://openalex.org/W7172380766","https://openalex.org/W7213397108"],"related_works":["https://openalex.org/W7172172326","https://openalex.org/W7215088263","https://openalex.org/W7162458638","https://openalex.org/W7213585152","https://openalex.org/W7172380766","https://openalex.org/W7214331945","https://openalex.org/W7127648241","https://openalex.org/W7164040202","https://openalex.org/W7216595918","https://openalex.org/W7221527653"],"abstract_inverted_index":{"World":[0],"Action":[1],"Models":[2],"(WAMs)":[3],"jointly":[4],"generate":[5],"video":[6],"and":[7,13,23,44,127,132,178,194,243],"robot":[8],"actions":[9,72],"through":[10,103],"iterative":[11],"diffusion":[12],"perform":[14],"strongly":[15],"in":[16,90,121,192,195],"robotic":[17],"manipulation.":[18],"However,":[19],"their":[20],"prohibitive":[21],"compute":[22],"memory":[24,220],"costs":[25],"pose":[26],"substantial":[27],"deployment":[28],"challenges.":[29],"Post-training":[30],"quantization":[31,66],"(PTQ)":[32],"can":[33],"reduce":[34],"these":[35],"costs,":[36],"but":[37],"existing":[38],"PTQ":[39],"methods":[40],"such":[41],"as":[42,138],"smoothing":[43],"rotation":[45],"are":[46],"insufficient":[47],"to":[48,134,148,182],"maintain":[49],"the":[50,71,73,79,100,115,129,157,163,168,173,199,214,219,222,230],"precision":[51],"of":[52,94,175,213,221],"action":[53,102,150,165,170],"generation.":[54],"To":[55],"overcome":[56],"this":[57],"limitation,":[58],"we":[59,77,155],"propose":[60],"Q-WAM,":[61],"a":[62,95,122,239,244],"new":[63],"4-bit":[64],"weight-activation":[65],"for":[67,162],"WAMs":[68],"that":[69,142,159],"preserves":[70],"model":[74],"generates.":[75],"Specifically,":[76],"introduce":[78],"\\textit{Action":[80],"Observability":[81],"Gramian":[82],"(AOG)},":[83],"which":[84,113],"measures":[85],"how":[86],"much":[87],"rounding":[88],"errors":[89],"each":[91,176],"weighted":[92],"combination":[93],"layer's":[96],"input":[97],"channels":[98],"change":[99],"final":[101],"all":[104],"denoising":[105],"steps.":[106],"We":[107,185],"also":[108],"develop":[109],"Action-Subspace":[110],"Protection":[111],"(ASP),":[112],"keeps":[114],"few":[116],"most":[117,161],"action-sensitive":[118],"channel":[119],"combinations":[120],"tiny":[123],"16-bit":[124,215],"low-rank":[125],"branch":[126],"quantizes":[128],"complementary":[130],"weights":[131],"activations":[133],"4":[135],"bits,":[136],"both":[137,191],"dense":[139],"matrix":[140],"multiplications":[141],"run":[143],"efficiently":[144],"on":[145,188],"GPUs.":[146],"Finally,":[147],"preserve":[149],"quality":[151],"with":[152],"minimal":[153],"overhead,":[154],"identify":[156],"experts":[158],"matter":[160],"generated":[164],"by":[166,225,234,253],"aggregating":[167],"AOG-derived":[169],"mass":[171],"across":[172],"layers":[174],"expert":[177],"apply":[179],"ASP":[180],"only":[181],"those":[183],"experts.":[184],"evaluate":[186],"Q-WAM":[187],"three":[189],"WAMs,":[190],"simulation":[193],"real-world":[196],"deployment.":[197],"On":[198,238],"RoboTwin":[200],"2.0":[201],"benchmark,":[202],"it":[203,248],"reaches":[204],"89.6--93.0\\%":[205],"average":[206],"success":[207,250],"rate,":[208],"within":[209],"1.1":[210],"percentage":[211,236,255],"points":[212],"models,":[216],"while":[217],"reducing":[218],"quantized":[223],"blocks":[224],"3.1--3.4$\\times$.":[226],"Our":[227],"method":[228],"outperforms":[229],"strongest":[231],"baseline,":[232],"SVDQuant,":[233],"2.5--8.7":[235],"points.":[237,256],"Unitree":[240],"G1":[241],"humanoid":[242],"bimanual":[245],"UR3":[246],"robot,":[247],"improves":[249],"over":[251],"SVDQuant":[252],"12.8-17.6":[254]},"counts_by_year":[],"updated_date":"2026-11-10T11:39:43.021295","created_date":"2026-09-30T00:00:00"}