{"id":"https://openalex.org/W7204475111","doi":"https://doi.org/10.48550/arxiv.2608.24987","title":"D$^3$-MOPD: Dynamic Domain ScheDuling for Efficient Multi-Teacher Distillation","display_name":"D$^3$-MOPD: Dynamic Domain ScheDuling for Efficient Multi-Teacher Distillation","publication_year":2026,"publication_date":"2026-08-25","ids":{"openalex":"https://openalex.org/W7204475111","doi":"https://doi.org/10.48550/arxiv.2608.24987"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.24987","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24987","pdf_url":"https://arxiv.org/pdf/2608.24987","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-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","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/2608.24987","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5126966045","display_name":"Zechen Sun","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sun, Zechen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148451462","display_name":"Zhiwei Zhang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Zhiwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123719284","display_name":"Fei Zhao","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Fei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125738571","display_name":"Juntao Li","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Juntao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148448611","display_name":"Mu Chuan","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chuan, Mu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5072289042","display_name":"Huayu Deng","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Deng, Huayu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5029301375","display_name":"Guojian Zhan","orcid":"https://orcid.org/0000-0002-1246-4860","observed_orcids":["https://orcid.org/0000-0002-1246-4860"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhan, Guojian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134169149","display_name":"Wenliang Chen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Wenliang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148420327","display_name":"Yao Hu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Yao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5148461408","display_name":"Min Zhang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Min","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":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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.8682000041007996,"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"}},"topics":[{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.8682000041007996,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.10459999740123749,"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/T10028","display_name":"Topic Modeling","score":0.004800000227987766,"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/multi-teacher-knowledge-distillation","display_name":"multi-teacher distillation","score":0.968999981880188},{"id":"https://openalex.org/keywords/on-policy-distillation","display_name":"on-policy distillation","score":0.9710000157356262},{"id":"https://openalex.org/keywords/reverse-kl-divergence","display_name":"reverse KL divergence","score":0.8489999771118164},{"id":"https://openalex.org/keywords/training-efficiency","display_name":"training efficiency","score":0.4650000035762787}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6567999720573425},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6473000049591064},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.5206999778747559},{"id":"https://openalex.org/C2776434776","wikidata":"https://www.wikidata.org/wiki/Q19246213","display_name":"Domain adaptation","level":3,"score":0.4307999908924103},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.4059000015258789},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.3765000104904175},{"id":"https://openalex.org/C2781395549","wikidata":"https://www.wikidata.org/wiki/Q4680762","display_name":"Adaptive sampling","level":3,"score":0.3614000082015991},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.35429999232292175},{"id":"https://openalex.org/C55416958","wikidata":"https://www.wikidata.org/wiki/Q6206757","display_name":"Job shop scheduling","level":3,"score":0.35339999198913574},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.34529998898506165},{"id":"https://openalex.org/C103824480","wikidata":"https://www.wikidata.org/wiki/Q185889","display_name":"Time domain","level":2,"score":0.32499998807907104},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3174000084400177},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3165000081062317},{"id":"https://openalex.org/C207390915","wikidata":"https://www.wikidata.org/wiki/Q1230525","display_name":"Divergence (linguistics)","level":2,"score":0.3158999979496002},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.310699999332428},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.29589998722076416},{"id":"https://openalex.org/C111873713","wikidata":"https://www.wikidata.org/wiki/Q1641413","display_name":"Job scheduler","level":3,"score":0.2906000018119812},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.27570000290870667},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.26820001006126404},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.26409998536109924}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.24987","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24987","pdf_url":"https://arxiv.org/pdf/2608.24987","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-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.24987","doi":"https://doi.org/10.48550/arxiv.2608.24987","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.24987","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-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2608.24987","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24987","pdf_url":"https://arxiv.org/pdf/2608.24987","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-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[{"score":0.5385000109672546,"display_name":"Quality education","id":"https://metadata.un.org/sdg/4"}],"sustainable_development_goals_aurora":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204475111.pdf","grobid_xml":"https://content.openalex.org/works/W7204475111.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7222992966","https://openalex.org/W7214796688","https://openalex.org/W7203946480","https://openalex.org/W7214759893","https://openalex.org/W7221507979","https://openalex.org/W7214617925","https://openalex.org/W7220978264","https://openalex.org/W7213995943","https://openalex.org/W7216515297","https://openalex.org/W7223099873"],"abstract_inverted_index":{"Multi-teacher":[0],"on-policy":[1],"distillation":[2],"(MOPD)":[3],"distills":[4],"several":[5],"domain-expert":[6,167],"teachers":[7,201],"into":[8],"a":[9,79,161],"single":[10],"student":[11,163],"by":[12],"minimizing":[13],"per-domain":[14,27,85],"reverse-KL":[15,86],"divergence":[16],"on":[17,61,202],"the":[18,26,33,52,84,94,101,123,129,143,156,173,185,199],"student's":[19],"own":[20],"rollouts.":[21],"Existing":[22],"approaches":[23],"typically":[24],"fix":[25],"data":[28],"mixture":[29,57,96],"before":[30],"training,":[31],"overlooking":[32],"fact":[34],"that":[35,82],"different":[36,41],"domains":[37,63,149],"converge":[38],"at":[39],"substantially":[40],"rates:":[42],"some":[43],"plateau":[44],"early":[45],"while":[46],"others":[47],"continue":[48],"to":[49,92,137,158],"improve":[50],"throughout":[51],"training":[53,91,102,131],"budget.":[54],"A":[55],"fixed":[56],"therefore":[58],"wastes":[59],"compute":[60],"fast-converging":[62],"and":[64,116,120,142,197],"undertrains":[65],"slower-converging":[66],"ones.":[67],"To":[68],"address":[69],"this,":[70],"we":[71],"propose":[72],"D$^3$-MOPD":[73,134,169],"(Dynamic":[74],"Domain":[75],"ScheDuling":[76],"for":[77,155,181],"MOPD),":[78],"zero-overhead":[80],"scheduler":[81,157],"repurposes":[83],"signal":[87],"already":[88],"produced":[89],"during":[90],"adapt":[93],"domain":[95,124],"online.":[97],"Running":[98],"asynchronously":[99],"outside":[100],"process,":[103],"an":[104,190],"off-process":[105],"watcher":[106],"periodically":[107],"tracks":[108],"each":[109],"domain's":[110],"KL":[111],"trajectory,":[112],"estimates":[113],"remaining":[114],"headroom":[115],"current":[117],"improvement":[118],"rate,":[119],"accordingly":[121],"adjusts":[122],"sampling":[125],"ratios":[126],"without":[127],"altering":[128],"core":[130],"loop.":[132],"Our":[133],"scales":[135],"naturally":[136],"arbitrary":[138],"numbers":[139],"of":[140,172,204],"domains,":[141],"expected":[144],"benefit":[145],"grows":[146],"as":[147],"more":[148,151],"introduce":[150],"diverse":[152],"convergence":[153],"patterns":[154],"exploit.":[159],"On":[160],"Qwen3.6-35B-A3B":[162],"distilled":[164],"from":[165],"four":[166],"teachers,":[168],"closes":[170],"97%":[171],"average":[174],"student-to-teacher":[175],"performance":[176,188],"gap,":[177],"compared":[178],"with":[179,189],"63%":[180],"vanilla":[182],"MOPD,":[183],"reaches":[184],"same":[186],"peak":[187],"approximately":[191],"3$\\times$":[192],"reduction":[193],"in":[194],"rollout":[195],"steps,":[196],"surpasses":[198],"specialist":[200],"three":[203],"seven":[205],"benchmarks.":[206]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-08-28T00:00:00"}