{"id":"https://openalex.org/W7169841323","doi":"https://doi.org/10.1145/3820441.3820458","title":"Semantic-Aware Loss Recovery for Cross-Datacenter Model Training","display_name":"Semantic-Aware Loss Recovery for Cross-Datacenter Model Training","publication_year":2026,"publication_date":"2026-07-21","ids":{"openalex":"https://openalex.org/W7169841323","doi":"https://doi.org/10.1145/3820441.3820458"},"language":null,"primary_location":{"id":"doi:10.1145/3820441.3820458","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3820441.3820458","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 10th Asia-Pacific Workshop on Networking","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3820441.3820458","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5141298825","display_name":"Xingjian Zhang","orcid":"https://orcid.org/0009-0008-4742-2629"},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xingjian Zhang","raw_affiliation_strings":["Beijing University of Posts and Telecommunications, Beijing, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0008-4742-2629","affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications, Beijing, Beijing, China","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134456262","display_name":"Yutong Zhao","orcid":"https://orcid.org/0009-0003-6029-3589"},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yutong Zhao","raw_affiliation_strings":["Beijing University of Posts and Telecommunications, Beijing, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0003-6029-3589","affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications, Beijing, Beijing, China","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073746201","display_name":"J R Liu","orcid":null},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiaxue Liu","raw_affiliation_strings":["Beijing University of Posts and Telecommunications, Beijing, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0008-2940-622X","affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications, Beijing, Beijing, China","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003744456","display_name":"Lizhuang Tan","orcid":"https://orcid.org/0000-0001-6826-4596"},"institutions":[{"id":"https://openalex.org/I4210158984","display_name":"National Supercomputing Center in Wuxi","ror":"https://ror.org/04ypjrs34","country_code":"CN","type":"other","lineage":["https://openalex.org/I4210158984"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lizhuang Tan","raw_affiliation_strings":["Shandong Computer Science Center (National Supercomputer Center in Ji'nan), Jinan, Shandong, China"],"raw_orcid":"https://orcid.org/0000-0001-6826-4596","affiliations":[{"raw_affiliation_string":"Shandong Computer Science Center (National Supercomputer Center in Ji'nan), Jinan, Shandong, China","institution_ids":["https://openalex.org/I4210158984"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054814598","display_name":"Shangguang Wang","orcid":"https://orcid.org/0000-0001-7245-1298"},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shangguang Wang","raw_affiliation_strings":["Beijing University of Posts and Telecommunications, Beijing, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0001-7245-1298","affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications, Beijing, Beijing, China","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5074291400","display_name":"Yiran Zhang","orcid":"https://orcid.org/0000-0002-3555-0155"},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yiran Zhang","raw_affiliation_strings":["Beijing University of Posts and Telecommunications, Beijing, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-3555-0155","affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications, Beijing, Beijing, China","institution_ids":["https://openalex.org/I139759216"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"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":"115","last_page":"121"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":null,"topics":[],"keywords":[{"id":"https://openalex.org/keywords/packet-loss","display_name":"Packet loss","score":0.6607999801635742},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.6333000063896179},{"id":"https://openalex.org/keywords/residual","display_name":"Residual","score":0.585099995136261},{"id":"https://openalex.org/keywords/convergence","display_name":"Convergence (economics)","score":0.5512999892234802},{"id":"https://openalex.org/keywords/data-loss","display_name":"Data loss","score":0.5291000008583069},{"id":"https://openalex.org/keywords/network-packet","display_name":"Network packet","score":0.4577000141143799},{"id":"https://openalex.org/keywords/scheme","display_name":"Scheme (mathematics)","score":0.42329999804496765}],"concepts":[{"id":"https://openalex.org/C54108766","wikidata":"https://www.wikidata.org/wiki/Q391064","display_name":"Packet loss","level":3,"score":0.6607999801635742},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6474000215530396},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.6333000063896179},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.585099995136261},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.5512999892234802},{"id":"https://openalex.org/C193519340","wikidata":"https://www.wikidata.org/wiki/Q891179","display_name":"Data loss","level":2,"score":0.5291000008583069},{"id":"https://openalex.org/C158379750","wikidata":"https://www.wikidata.org/wiki/Q214111","display_name":"Network packet","level":2,"score":0.4577000141143799},{"id":"https://openalex.org/C44154836","wikidata":"https://www.wikidata.org/wiki/Q45045","display_name":"Simulation","level":1,"score":0.4553999900817871},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.4404999911785126},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.42329999804496765},{"id":"https://openalex.org/C130795937","wikidata":"https://www.wikidata.org/wiki/Q2561570","display_name":"Remote direct memory access","level":2,"score":0.37940001487731934},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.37599998712539673},{"id":"https://openalex.org/C38349280","wikidata":"https://www.wikidata.org/wiki/Q1434290","display_name":"Flow (mathematics)","level":2,"score":0.3418000042438507},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.32820001244544983},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.29420000314712524},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.26460000872612},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.2623000144958496},{"id":"https://openalex.org/C2779764123","wikidata":"https://www.wikidata.org/wiki/Q7972881","display_name":"Wasting","level":2,"score":0.25769999623298645}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3820441.3820458","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3820441.3820458","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 10th Asia-Pacific Workshop on Networking","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3820441.3820458","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3820441.3820458","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 10th Asia-Pacific Workshop on Networking","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G6778685876","display_name":null,"funder_award_id":"2024YFB2907000","funder_id":"https://openalex.org/F4320336806","funder_display_name":"National Natural Science Foundation of China - State Grid Corporation Joint Fund for Smart Grid"}],"funders":[{"id":"https://openalex.org/F4320336806","display_name":"National Natural Science Foundation of China - State Grid Corporation Joint Fund for Smart Grid","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":6,"referenced_works":["https://openalex.org/W2064675550","https://openalex.org/W2809353470","https://openalex.org/W4394945032","https://openalex.org/W4402896955","https://openalex.org/W4404239922","https://openalex.org/W4416203763"],"related_works":[],"abstract_inverted_index":{"Cross-datacenter":[0],"distributed":[1],"model":[2,46],"training":[3],"is":[4],"increasingly":[5],"important,":[6],"but":[7],"WAN":[8],"packet":[9],"loss":[10,41],"and":[11,25,80,96,105],"long":[12],"propagation":[13],"delays":[14],"make":[15],"retransmission-based":[16],"recovery":[17,42,104],"expensive.":[18],"Existing":[19],"RDMA":[20],"FEC":[21],"schemes":[22],"are":[23],"application-unaware":[24],"uniformly":[26],"protect":[27],"all":[28],"packets,":[29,66],"wasting":[30],"inter-datacenter":[31],"bandwidth.":[32],"In":[33],"this":[34],"paper,":[35],"we":[36],"present":[37],"SMART,":[38],"a":[39],"semantic-aware":[40],"scheme":[43],"for":[44],"cross-datacenter":[45],"training.":[47],"SMART":[48,93],"distinguishes":[49],"data-parallel":[50],"(DP)":[51],"gradients":[52],"from":[53],"pipeline-parallel":[54],"(PP)":[55],"activations":[56],"using":[57],"lightweight":[58],"message-level":[59],"tags;":[60],"it":[61],"selectively":[62],"protects":[63],"high-priority":[64],"DP":[65,84],"gives":[67],"earlier":[68],"PP":[69,78],"traffic":[70],"stronger":[71],"FEC,":[72,107],"zero-fills":[73],"only":[74],"small":[75],"profiled":[76],"residual":[77],"loss,":[79],"compensates":[81],"tolerated":[82],"low-priority":[83],"losses.":[85],"Simulation":[86],"experiments":[87],"with":[88,102],"Qwen3-3B-derived":[89],"workloads":[90],"show":[91],"that":[92],"reduces":[94],"average":[95],"P99":[97],"flow":[98],"completion":[99],"times":[100],"compared":[101],"retransmission-only":[103],"uniform":[106],"while":[108],"preserving":[109],"convergence":[110],"close":[111],"to":[112],"the":[113],"no-loss":[114],"baseline.":[115]},"counts_by_year":[],"updated_date":"2026-07-23T08:03:31.855105","created_date":"2026-07-22T00:00:00"}
