{"id":"https://openalex.org/W4416749079","doi":"https://doi.org/10.1109/iros60139.2025.11247780","title":"AdaToken-3D: Dynamic Spatial Gating for Efficient 3D Large Multimodal-Models Reasoning","display_name":"AdaToken-3D: Dynamic Spatial Gating for Efficient 3D Large Multimodal-Models Reasoning","publication_year":2025,"publication_date":"2025-10-19","ids":{"openalex":"https://openalex.org/W4416749079","doi":"https://doi.org/10.1109/iros60139.2025.11247780"},"language":null,"primary_location":{"id":"doi:10.1109/iros60139.2025.11247780","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iros60139.2025.11247780","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100324071","display_name":"Kai Zhang","orcid":"https://orcid.org/0009-0005-6803-7533"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kai Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5056664840","display_name":"Xingyu Chen","orcid":"https://orcid.org/0000-0002-1989-4935"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xingyu Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5100455676","display_name":"Xiaofeng Zhang","orcid":"https://orcid.org/0000-0003-0972-8842"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiaofeng Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"16702","last_page":"16709"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.7860999703407288,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.7860999703407288,"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":0.06920000165700912,"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":0.016200000420212746,"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/security-token","display_name":"Security token","score":0.6291000247001648},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6154999732971191},{"id":"https://openalex.org/keywords/redundancy","display_name":"Redundancy (engineering)","score":0.5963000059127808},{"id":"https://openalex.org/keywords/spatial-analysis","display_name":"Spatial analysis","score":0.5076000094413757},{"id":"https://openalex.org/keywords/focus","display_name":"Focus (optics)","score":0.5065000057220459},{"id":"https://openalex.org/keywords/spatial-intelligence","display_name":"Spatial intelligence","score":0.3968000113964081},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.38839998841285706},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.38690000772476196},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.3652999997138977}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8051999807357788},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.6291000247001648},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6154999732971191},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.5963000059127808},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5494999885559082},{"id":"https://openalex.org/C159620131","wikidata":"https://www.wikidata.org/wiki/Q1938983","display_name":"Spatial analysis","level":2,"score":0.5076000094413757},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.5065000057220459},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.42100000381469727},{"id":"https://openalex.org/C155911833","wikidata":"https://www.wikidata.org/wiki/Q3817354","display_name":"Spatial intelligence","level":2,"score":0.3968000113964081},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.38839998841285706},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.38690000772476196},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.3652999997138977},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.3450999855995178},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.329800009727478},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.3237999975681305},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.31700000166893005},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.31279999017715454},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.30720001459121704},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.2874999940395355},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.27410000562667847},{"id":"https://openalex.org/C27511587","wikidata":"https://www.wikidata.org/wiki/Q2178623","display_name":"Spatial relation","level":2,"score":0.2678000032901764},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.2646999955177307},{"id":"https://openalex.org/C64754055","wikidata":"https://www.wikidata.org/wiki/Q7574053","display_name":"Spatial contextual awareness","level":2,"score":0.26330000162124634},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.26030001044273376},{"id":"https://openalex.org/C2776650193","wikidata":"https://www.wikidata.org/wiki/Q264661","display_name":"Obstacle","level":2,"score":0.25609999895095825}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/iros60139.2025.11247780","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iros60139.2025.11247780","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":12,"referenced_works":["https://openalex.org/W2594519801","https://openalex.org/W3016211260","https://openalex.org/W3107521863","https://openalex.org/W3182910454","https://openalex.org/W4226376247","https://openalex.org/W4312290555","https://openalex.org/W4402716423","https://openalex.org/W4403081466","https://openalex.org/W4403939369","https://openalex.org/W4404003268","https://openalex.org/W4411119489","https://openalex.org/W4411120330"],"related_works":[],"abstract_inverted_index":{"Large":[0],"Multimodal":[1],"Models":[2],"(LMMs)":[3],"have":[4],"become":[5],"a":[6],"pivotal":[7],"research":[8],"focus":[9],"in":[10,16,142],"deep":[11],"learning,":[12],"demonstrating":[13],"remarkable":[14],"capabilities":[15],"3D":[17,22,49,96,173],"scene":[18],"understanding.":[19],"However,":[20],"current":[21],"LMMs":[23,50],"employing":[24],"thousands":[25],"of":[26,158],"spatial":[27,61,75,85,144,159],"tokens":[28,62,83,160],"for":[29,171],"multimodal":[30,143,174],"reasoning":[31],"suffer":[32],"from":[33],"critical":[34],"inefficiencies:":[35],"excessive":[36],"computational":[37],"overhead":[38],"and":[39,63,124],"redundant":[40,82],"information":[41,102,145],"flows.":[42],"Unlike":[43],"2D":[44],"VLMs":[45],"processing":[46],"single":[47],"images,":[48],"exhibit":[51],"inherent":[52],"architectural":[53],"redundancy":[54,140],"due":[55],"to":[56,94,164],"the":[57,165],"heterogeneous":[58],"mechanisms":[59],"between":[60],"visual":[64],"tokens.":[65],"To":[66],"address":[67],"this":[68,136],"challenge,":[69],"we":[70],"propose":[71],"AdaToken-3D,":[72],"an":[73],"adaptive":[74],"token":[76,149],"optimization":[77],"framework":[78],"that":[79,117,155],"dynamically":[80],"prunes":[81],"through":[84,147],"contribution":[86],"analysis.":[87,151],"Our":[88,152],"method":[89],"automatically":[90],"tailors":[91],"pruning":[92],"strategies":[93],"different":[95],"LMM":[97],"architectures":[98],"by":[99],"quantifying":[100],"token-level":[101],"flows":[103,146],"via":[104],"attention":[105],"pattern":[106],"mining.":[107],"Extensive":[108],"experiments":[109],"on":[110],"LLaVA-3D":[111],"(a":[112],"7B":[113],"parameter":[114],"3D-LMM)":[115],"demonstrate":[116],"AdaToken-3D":[118],"achieves":[119],"21%":[120],"faster":[121],"inference":[122],"speed":[123],"63%":[125],"FLOPs":[126],"reduction":[127],"while":[128],"maintaining":[129],"original":[130],"task":[131],"accuracy.":[132],"Beyond":[133],"efficiency":[134],"gains,":[135],"work":[137],"systematically":[138],"investigates":[139],"patterns":[141],"quantitative":[148],"interaction":[150],"findings":[153],"reveal":[154],"over":[156],"60%":[157],"contribute":[161],"minimally":[162],"(<5%)":[163],"final":[166],"predictions,":[167],"establishing":[168],"theoretical":[169],"foundations":[170],"efficient":[172],"learning.":[175]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-11-28T00:00:00"}