{"id":"https://openalex.org/W4416214699","doi":"https://doi.org/10.1109/lra.2025.3632751","title":"Mini Diffuser: Fast Multi-Task Diffusion Policy Training Using Two-Level Mini-Batches","display_name":"Mini Diffuser: Fast Multi-Task Diffusion Policy Training Using Two-Level Mini-Batches","publication_year":2025,"publication_date":"2025-11-14","ids":{"openalex":"https://openalex.org/W4416214699","doi":"https://doi.org/10.1109/lra.2025.3632751"},"language":"en","primary_location":{"id":"doi:10.1109/lra.2025.3632751","is_oa":false,"landing_page_url":"https://doi.org/10.1109/lra.2025.3632751","pdf_url":null,"source":{"id":"https://openalex.org/S4210169774","display_name":"IEEE Robotics and Automation Letters","issn_l":"2377-3766","issn":["2377-3766","2377-3774"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Robotics and Automation Letters","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://lirias.kuleuven.be/retrieve/8bb84ea3-8330-43d2-94d6-bfce513492fd","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5056518878","display_name":"Yutong Hu","orcid":"https://orcid.org/0000-0002-6030-3113"},"institutions":[{"id":"https://openalex.org/I99464096","display_name":"KU Leuven","ror":"https://ror.org/05f950310","country_code":"BE","type":"education","lineage":["https://openalex.org/I99464096"]}],"countries":["BE"],"is_corresponding":false,"raw_author_name":"Yutong Hu","raw_affiliation_strings":["Department Mechanical Engineering, Research unit Robotics, Automation and Mechatronics, KU Leuven, Leuven, Belgium","Dept. Mechanical Engineering, Research unit Robotics, Automation and Mechatronics, KU Leuven, Leuven, Belgium"],"raw_orcid":"https://orcid.org/0000-0002-6030-3113","affiliations":[{"raw_affiliation_string":"Department Mechanical Engineering, Research unit Robotics, Automation and Mechatronics, KU Leuven, Leuven, Belgium","institution_ids":["https://openalex.org/I99464096"]},{"raw_affiliation_string":"Dept. Mechanical Engineering, Research unit Robotics, Automation and Mechatronics, KU Leuven, Leuven, Belgium","institution_ids":["https://openalex.org/I99464096"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048813179","display_name":"Pinhao Song","orcid":"https://orcid.org/0000-0002-3950-0403"},"institutions":[{"id":"https://openalex.org/I99464096","display_name":"KU Leuven","ror":"https://ror.org/05f950310","country_code":"BE","type":"education","lineage":["https://openalex.org/I99464096"]}],"countries":["BE"],"is_corresponding":false,"raw_author_name":"Pinhao Song","raw_affiliation_strings":["Department Mechanical Engineering, Research unit Robotics, Automation and Mechatronics, KU Leuven, Leuven, Belgium","Dept. Mechanical Engineering, Research unit Robotics, Automation and Mechatronics, KU Leuven, Leuven, Belgium"],"raw_orcid":"https://orcid.org/0000-0002-3950-0403","affiliations":[{"raw_affiliation_string":"Department Mechanical Engineering, Research unit Robotics, Automation and Mechatronics, KU Leuven, Leuven, Belgium","institution_ids":["https://openalex.org/I99464096"]},{"raw_affiliation_string":"Dept. Mechanical Engineering, Research unit Robotics, Automation and Mechatronics, KU Leuven, Leuven, Belgium","institution_ids":["https://openalex.org/I99464096"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113150806","display_name":"Kehan Wen","orcid":"https://orcid.org/0000-0003-0411-4489"},"institutions":[{"id":"https://openalex.org/I35440088","display_name":"ETH Zurich","ror":"https://ror.org/05a28rw58","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I35440088"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Kehan Wen","raw_affiliation_strings":["Department of Information Technology and Electrical Engineering, ETH Z&#x00FC;rich, Zurich, Switzerland","Dept. Electrical Engineering, ETH Z&#x00FC;rich, Zurich, Switzerland"],"raw_orcid":"https://orcid.org/0000-0003-0411-4489","affiliations":[{"raw_affiliation_string":"Department of Information Technology and Electrical Engineering, ETH Z&#x00FC;rich, Zurich, Switzerland","institution_ids":["https://openalex.org/I35440088"]},{"raw_affiliation_string":"Dept. Electrical Engineering, ETH Z&#x00FC;rich, Zurich, Switzerland","institution_ids":["https://openalex.org/I35440088"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5049896600","display_name":"Renaud Detry","orcid":"https://orcid.org/0000-0003-0597-1167"},"institutions":[{"id":"https://openalex.org/I99464096","display_name":"KU Leuven","ror":"https://ror.org/05f950310","country_code":"BE","type":"education","lineage":["https://openalex.org/I99464096"]}],"countries":["BE"],"is_corresponding":false,"raw_author_name":"Renaud Detry","raw_affiliation_strings":["Department Electrical Engineering, Research unit Processing Speech and Images, KU Leuven, Leuven, Belgium","Dept. Mechanical Engineering, Research unit Robotics, Automation and Mechatronics, KU Leuven, Leuven, Belgium"],"raw_orcid":"https://orcid.org/0000-0003-0597-1167","affiliations":[{"raw_affiliation_string":"Department Electrical Engineering, Research unit Processing Speech and Images, KU Leuven, Leuven, Belgium","institution_ids":["https://openalex.org/I99464096"]},{"raw_affiliation_string":"Dept. Mechanical Engineering, Research unit Robotics, Automation and Mechatronics, KU Leuven, Leuven, Belgium","institution_ids":["https://openalex.org/I99464096"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":{"value":2645,"currency":"USD","value_usd":2645},"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.23561996,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"11","issue":"1","first_page":"938","last_page":"945"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.41269999742507935,"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":0.41269999742507935,"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":0.2401999980211258,"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":0.11110000312328339,"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/curse-of-dimensionality","display_name":"Curse of dimensionality","score":0.605400025844574},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.545799970626831},{"id":"https://openalex.org/keywords/action","display_name":"Action (physics)","score":0.44369998574256897},{"id":"https://openalex.org/keywords/diffusion","display_name":"Diffusion","score":0.384799987077713},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.38269999623298645},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.365200012922287},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.33239999413490295},{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.33239999413490295}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6506999731063843},{"id":"https://openalex.org/C111030470","wikidata":"https://www.wikidata.org/wiki/Q1430460","display_name":"Curse of dimensionality","level":2,"score":0.605400025844574},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5785999894142151},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.545799970626831},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.44369998574256897},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3896999955177307},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.384799987077713},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.38269999623298645},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.365200012922287},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.33239999413490295},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.33239999413490295},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.33219999074935913},{"id":"https://openalex.org/C55128770","wikidata":"https://www.wikidata.org/wiki/Q5275440","display_name":"Diffusion map","level":4,"score":0.33219999074935913},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3197999894618988},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.30239999294281006},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.2759999930858612},{"id":"https://openalex.org/C9417928","wikidata":"https://www.wikidata.org/wiki/Q1070689","display_name":"Image processing","level":3,"score":0.2718999981880188},{"id":"https://openalex.org/C203504353","wikidata":"https://www.wikidata.org/wiki/Q4765461","display_name":"Anisotropic diffusion","level":3,"score":0.26899999380111694},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.265500009059906},{"id":"https://openalex.org/C2777042071","wikidata":"https://www.wikidata.org/wiki/Q6509304","display_name":"Leakage (economics)","level":2,"score":0.25780001282691956}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/lra.2025.3632751","is_oa":false,"landing_page_url":"https://doi.org/10.1109/lra.2025.3632751","pdf_url":null,"source":{"id":"https://openalex.org/S4210169774","display_name":"IEEE Robotics and Automation Letters","issn_l":"2377-3766","issn":["2377-3766","2377-3774"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Robotics and Automation Letters","raw_type":"journal-article"},{"id":"pmh:oai:lirias2repo.kuleuven.be:20.500.12942/777534","is_oa":true,"landing_page_url":"https://lirias.kuleuven.be/handle/20.500.12942/777534","pdf_url":"https://lirias.kuleuven.be/retrieve/8bb84ea3-8330-43d2-94d6-bfce513492fd","source":{"id":"https://openalex.org/S7407055369","display_name":"Lirias","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"acceptedVersion","is_accepted":true,"is_published":false,"raw_source_name":"IEEE Robotics and Automation Letters, vol. 11 (1), (938-945)","raw_type":"info:eu-repo/semantics/article"}],"best_oa_location":{"id":"pmh:oai:lirias2repo.kuleuven.be:20.500.12942/777534","is_oa":true,"landing_page_url":"https://lirias.kuleuven.be/handle/20.500.12942/777534","pdf_url":"https://lirias.kuleuven.be/retrieve/8bb84ea3-8330-43d2-94d6-bfce513492fd","source":{"id":"https://openalex.org/S7407055369","display_name":"Lirias","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"acceptedVersion","is_accepted":true,"is_published":false,"raw_source_name":"IEEE Robotics and Automation Letters, vol. 11 (1), (938-945)","raw_type":"info:eu-repo/semantics/article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G379021153","display_name":null,"funder_award_id":"C14/22/134","funder_id":"https://openalex.org/F4320322308","funder_display_name":"KU Leuven"}],"funders":[{"id":"https://openalex.org/F4320319128","display_name":"Vlaams Supercomputer Centrum","ror":null},{"id":"https://openalex.org/F4320320300","display_name":"European Commission","ror":"https://ror.org/00k4n6c32"},{"id":"https://openalex.org/F4320321652","display_name":"Eidgen\u00f6ssische Technische Hochschule Z\u00fcrich","ror":"https://ror.org/05a28rw58"},{"id":"https://openalex.org/F4320321730","display_name":"Fonds Wetenschappelijk Onderzoek","ror":"https://ror.org/03qtxy027"},{"id":"https://openalex.org/F4320322308","display_name":"KU Leuven","ror":"https://ror.org/05f950310"},{"id":"https://openalex.org/F4320327336","display_name":"Vlaamse regering","ror":null}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416214699.pdf","grobid_xml":"https://content.openalex.org/works/W4416214699.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"We":[0],"present":[1],"a":[2,27],"method":[3],"that":[4,39,108,150],"reduces,":[5],"by":[6,75],"an":[7],"order":[8],"of":[9,56,89,125,128,137,143,156],"magnitude,":[10],"the":[11,35,45,54,57,64,90,105,126,138,144,153,160],"time":[12,140],"and":[13,34,62,141,167,182],"memory":[14],"needed":[15],"to":[16,104,162],"train":[17],"multi-task":[18,130],"vision-language":[19,86],"robotic":[20],"diffusion":[21,33,37,106,131],"policies.":[22],"This":[23],"improvement":[24],"arises":[25],"from":[26],"previously":[28],"underexplored":[29],"distinction":[30],"between":[31],"action":[32,52,82,165],"image":[36,43,65],"techniques":[38],"inspired":[40],"it:":[41],"In":[42,119],"generation,":[44,53],"target":[46,58],"is":[47,59,67],"high-dimensional.":[48,68],"By":[49],"contrast,":[50],"in":[51],"dimensionality":[55],"comparatively":[60],"small,":[61],"only":[63,135],"condition":[66],"Our":[69],"approach,Mini":[70],"Diffuser,":[71],"exploits":[72],"this":[73,97],"asymmetry":[74],"introducingtwo-level":[76],"minibatching,":[77],"which":[78],"pairs":[79],"multiple":[80],"noised":[81],"samples":[83,113],"with":[84,180],"each":[85],"condition,":[87],"instead":[88],"conventional":[91],"one-to-one":[92],"sampling":[93],"strategy.":[94],"To":[95],"support":[96],"batching":[98],"scheme,":[99],"we":[100],"introduce":[101],"architectural":[102],"adaptations":[103],"transformer":[107],"prevent":[109],"information":[110],"leakage":[111],"across":[112],"while":[114,133],"maintaining":[115],"full":[116],"conditioning":[117],"access.":[118],"RLBench":[120],"simulations,":[121],"Mini-Diffuser":[122,151],"achieves":[123],"95%":[124],"performance":[127],"state-of-the-art":[129],"policies,":[132,158],"using":[134],"5%":[136],"training":[139,183],"7%":[142],"memory.":[145],"Real-world":[146],"experiments":[147],"further":[148],"validate":[149],"preserves":[152],"key":[154],"strengths":[155],"diffusion-based":[157],"including":[159],"ability":[161],"model":[163],"multimodal":[164],"distributions":[166],"produce":[168],"behavior":[169],"conditioned":[170],"on":[171],"diverse":[172],"perceptual":[173],"inputs.":[174],"Code":[175],"available":[176],"at":[177],"mini-diffuse-actor.github.io":[178],"along":[179],"videos":[181],"logs.":[184]},"counts_by_year":[],"updated_date":"2026-08-21T09:56:20.448147","created_date":"2025-11-14T00:00:00"}
