{"id":"https://openalex.org/W7134818722","doi":"https://doi.org/10.48550/arxiv.2603.06952","title":"Not All Neighbors Matter: Understanding the Impact of Graph Sparsification on GNN Pipelines","display_name":"Not All Neighbors Matter: Understanding the Impact of Graph Sparsification on GNN Pipelines","publication_year":2026,"publication_date":"2026-03-07","ids":{"openalex":"https://openalex.org/W7134818722","doi":"https://doi.org/10.48550/arxiv.2603.06952"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2603.06952","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","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":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128671768","display_name":"Yuhang Song","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Song, Yuhang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5118706450","display_name":"Naima Abrar Shami","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shami, Naima Abrar","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5050085800","display_name":"Romaric Duvignau","orcid":"https://orcid.org/0000-0003-1268-9311"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Duvignau, Romaric","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5068567119","display_name":"Vasiliki Kalavri","orcid":"https://orcid.org/0000-0001-8219-4862"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kalavri, Vasiliki","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5128671768"],"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.9854000210762024,"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.9854000210762024,"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/T12292","display_name":"Graph Theory and Algorithms","score":0.00430000014603138,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.0012000000569969416,"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/inference","display_name":"Inference","score":0.5856000185012817},{"id":"https://openalex.org/keywords/graph","display_name":"Graph","score":0.5033000111579895},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.4359999895095825},{"id":"https://openalex.org/keywords/node","display_name":"Node (physics)","score":0.36579999327659607},{"id":"https://openalex.org/keywords/pipeline-transport","display_name":"Pipeline transport","score":0.36070001125335693},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.3100999891757965}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.796500027179718},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5856000185012817},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.5033000111579895},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.46810001134872437},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.4359999895095825},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.42489999532699585},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.421999990940094},{"id":"https://openalex.org/C62611344","wikidata":"https://www.wikidata.org/wiki/Q1062658","display_name":"Node (physics)","level":2,"score":0.36579999327659607},{"id":"https://openalex.org/C175309249","wikidata":"https://www.wikidata.org/wiki/Q725864","display_name":"Pipeline transport","level":2,"score":0.36070001125335693},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3368000090122223},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.3100999891757965},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.30230000615119934},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.301800012588501},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.28780001401901245},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.28679999709129333},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.28529998660087585},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2711000144481659},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.2639000117778778}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2603.06952","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2603.06952","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.06952","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":"pmh:doi:10.48550/arxiv.2603.06952","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","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":{"As":[0,137],"graphs":[1],"scale":[2],"to":[3,34,65,76],"billions":[4],"of":[5,18,97,107,116,145,196],"nodes":[6],"and":[7,28,43,105,119,164],"edges,":[8],"graph":[9,57,180],"Machine":[10],"Learning":[11],"workloads":[12],"are":[13],"constrained":[14],"by":[15,149,181],"the":[16,46,103,112,143,146,152,170,178,193],"cost":[17],"multi-hop":[19],"traversals":[20],"over":[21],"exponentially":[22],"growing":[23],"neighborhoods.":[24],"While":[25],"various":[26],"system-level":[27],"algorithmic":[29],"optimizations":[30],"have":[31],"been":[32],"proposed":[33],"accelerate":[35],"Graph":[36],"Neural":[37],"Network":[38],"(GNN)":[39],"pipelines,":[40],"data":[41],"management":[42],"movement":[44],"remain":[45],"primary":[47],"bottlenecks":[48,79],"at":[49],"scale.":[50],"In":[51],"this":[52],"paper,":[53],"we":[54,190],"explore":[55],"whether":[56],"sparsification,":[58],"a":[59,72,185],"well-established":[60],"technique":[61],"that":[62,93,169,192],"reduces":[63],"edges":[64],"create":[66],"sparser":[67],"neighborhoods,":[68],"can":[69],"serve":[70],"as":[71],"lightweight":[73],"pre-processing":[74],"step":[75],"address":[77],"these":[78],"while":[80],"preserving":[81],"accuracy":[82,106,144,187],"on":[83,121,151,177],"node":[84],"classification":[85],"tasks.":[86],"We":[87,110],"develop":[88],"an":[89,138],"extensible":[90],"experimental":[91],"framework":[92],"enables":[94],"systematic":[95],"evaluation":[96],"how":[98],"different":[99],"sparsification":[100,129,141,197],"methods":[101],"affect":[102],"performance":[104,176],"GNN":[108,117],"models.":[109],"conduct":[111],"first":[113],"comprehensive":[114],"study":[115],"training":[118,163],"inference":[120],"sparsified":[122],"graphs,":[123],"revealing":[124],"several":[125],"key":[126],"findings.":[127],"First,":[128],"often":[130],"preserves":[131],"or":[132],"even":[133],"improves":[134,173],"predictive":[135],"performance.":[136],"example,":[139],"random":[140],"raises":[142],"GAT":[147],"model":[148,174],"6.8%":[150],"PubMed":[153],"graph.":[154],"Second,":[155],"benefits":[156],"increase":[157],"with":[158,183],"scale,":[159],"substantially":[160],"accelerating":[161],"both":[162],"inference.":[165],"Our":[166],"results":[167],"show":[168],"K-Neighbor":[171],"sparsifier":[172],"serving":[175],"Products":[179],"11.7x":[182],"only":[184],"0.7%":[186],"drop.":[188],"Importantly,":[189],"find":[191],"computational":[194],"overhead":[195],"is":[198],"quickly":[199],"amortized,":[200],"making":[201],"it":[202],"practical":[203],"for":[204],"very":[205],"large":[206],"graphs.":[207]},"counts_by_year":[],"updated_date":"2026-04-04T16:13:02.066488","created_date":"2026-03-11T00:00:00"}