{"id":"https://openalex.org/W7163027700","doi":"https://doi.org/10.48550/arxiv.2605.31164","title":"D$^3$: Dynamic Directional Graph-Constrained Data Scheduling for LLM Training","display_name":"D$^3$: Dynamic Directional Graph-Constrained Data Scheduling for LLM Training","publication_year":2026,"publication_date":"2026-05-29","ids":{"openalex":"https://openalex.org/W7163027700","doi":"https://doi.org/10.48550/arxiv.2605.31164"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.31164","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.31164","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.31164","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5123573670","display_name":"Yuanjian Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Yuanjian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123558253","display_name":"Jianing Hao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hao, Jianing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137516640","display_name":"Guang Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Guang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5137609750","display_name":"Zhong Li","orcid":"https://orcid.org/0000-0001-6087-6854"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Zhong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.18799999356269836,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.18799999356269836,"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/T11719","display_name":"Data Quality and Management","score":0.15320000052452087,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T11273","display_name":"Advanced Graph Neural Networks","score":0.10080000013113022,"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/scheduling","display_name":"Scheduling (production processes)","score":0.6211000084877014},{"id":"https://openalex.org/keywords/dynamic-priority-scheduling","display_name":"Dynamic priority scheduling","score":0.4846999943256378},{"id":"https://openalex.org/keywords/graph","display_name":"Graph","score":0.4187999963760376},{"id":"https://openalex.org/keywords/fair-share-scheduling","display_name":"Fair-share scheduling","score":0.3889999985694885},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.38659998774528503},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.3817000091075897},{"id":"https://openalex.org/keywords/data-modeling","display_name":"Data modeling","score":0.3571999967098236},{"id":"https://openalex.org/keywords/source-code","display_name":"Source code","score":0.337799996137619}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8108999729156494},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6211000084877014},{"id":"https://openalex.org/C107568181","wikidata":"https://www.wikidata.org/wiki/Q5319000","display_name":"Dynamic priority scheduling","level":3,"score":0.4846999943256378},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.4187999963760376},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4092999994754791},{"id":"https://openalex.org/C31689143","wikidata":"https://www.wikidata.org/wiki/Q733809","display_name":"Fair-share scheduling","level":3,"score":0.3889999985694885},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.38659998774528503},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.38370001316070557},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3817000091075897},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.37929999828338623},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.3571999967098236},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.337799996137619},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3352000117301941},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.30799999833106995},{"id":"https://openalex.org/C55416958","wikidata":"https://www.wikidata.org/wiki/Q6206757","display_name":"Job shop scheduling","level":3,"score":0.295199990272522},{"id":"https://openalex.org/C119948110","wikidata":"https://www.wikidata.org/wiki/Q7858726","display_name":"Two-level scheduling","level":4,"score":0.29109999537467957},{"id":"https://openalex.org/C88468194","wikidata":"https://www.wikidata.org/wiki/Q1172416","display_name":"Data-flow analysis","level":3,"score":0.2888999879360199},{"id":"https://openalex.org/C3020493868","wikidata":"https://www.wikidata.org/wiki/Q55631277","display_name":"Real world data","level":2,"score":0.2712000012397766},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.266400009393692},{"id":"https://openalex.org/C197298091","wikidata":"https://www.wikidata.org/wiki/Q5318963","display_name":"Dynamic data","level":2,"score":0.263700008392334},{"id":"https://openalex.org/C133462117","wikidata":"https://www.wikidata.org/wiki/Q4929239","display_name":"Data collection","level":2,"score":0.26159998774528503},{"id":"https://openalex.org/C2780440489","wikidata":"https://www.wikidata.org/wiki/Q5227278","display_name":"Data-driven","level":2,"score":0.25949999690055847},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.257999986410141},{"id":"https://openalex.org/C127456818","wikidata":"https://www.wikidata.org/wiki/Q238879","display_name":"Rate-monotonic scheduling","level":4,"score":0.25369998812675476},{"id":"https://openalex.org/C112866106","wikidata":"https://www.wikidata.org/wiki/Q267053","display_name":"Lottery scheduling","level":5,"score":0.2535000145435333},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2533999979496002}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.31164","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.31164","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.31164","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.31164","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Training":[0],"data":[1,16,27,49,124,144],"plays":[2],"a":[3,81,96,108,170],"central":[4],"role":[5],"in":[6],"large":[7],"language":[8],"models":[9],"(LLMs)":[10],"optimization,":[11],"motivating":[12],"extensive":[13],"research":[14],"on":[15,23,55],"scheduling":[17,86,145],"strategies.":[18],"Most":[19],"existing":[20,143],"approaches":[21],"concentrate":[22],"adjusting":[24],"the":[25,31,59,90,117,123,127,165,176],"overall":[26],"distribution":[28],"but":[29],"neglect":[30],"underlying":[32],"interactions":[33,43,92],"between":[34],"samples":[35,50],"during":[36],"training.":[37,132],"However,":[38],"we":[39,64,78],"argue":[40],"that":[41,122,163],"such":[42],"cannot":[44],"be":[45],"overlooked,":[46],"as":[47,95],"real-world":[48],"frequently":[51],"exhibit":[52],"directional":[53],"influences":[54],"each":[56],"other,":[57],"making":[58],"training":[60,118],"order":[61],"crucial.":[62],"Intuitively,":[63],"can":[65],"prioritize":[66],"train-units":[67,94],"with":[68],"greater":[69],"influence":[70,98],"to":[71,115],"improves":[72],"learning":[73],"efficiency.":[74],"In":[75],"this":[76,113],"work,":[77],"propose":[79],"$D^3$,":[80],"Dynamic":[82],"Directional":[83],"graph-constrained":[84],"Data":[85],"framework.":[87],"$D^3$":[88,156],"formulates":[89],"complex":[91],"among":[93],"dynamic":[97],"graph,":[99],"where":[100],"edges":[101],"represent":[102],"loss-based":[103],"dependencies.":[104],"It":[105],"then":[106],"solves":[107],"constrained":[109],"optimization":[110],"problem":[111],"over":[112,142],"graph":[114],"derive":[116],"order,":[119],"which":[120],"ensures":[121],"sequence":[125],"respects":[126],"evolving":[128],"information":[129],"flow":[130],"throughout":[131],"Our":[133],"approach":[134],"is":[135,178],"theoretically":[136],"motivated":[137],"and":[138,150],"yields":[139],"consistent":[140],"improvements":[141],"methods":[146],"across":[147],"both":[148],"pre-training":[149],"post-training":[151],"phases.":[152],"Furthermore,":[153],"for":[154],"scalability,":[155],"also":[157],"employs":[158],"an":[159],"efficient":[160],"approximation":[161],"algorithm":[162],"keeps":[164],"additional":[166],"computational":[167],"overhead":[168],"within":[169],"manageable":[171],"range.":[172],"For":[173],"future":[174],"research,":[175],"code":[177],"available":[179],"at":[180],"https://github.com/xuyj233/D3.":[181]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-06-02T00:00:00"}