{"id":"https://openalex.org/W7225808538","doi":"https://doi.org/10.21203/rs.3.rs-11136263/v1","title":"SOFIA: Quantifying Sampling-Induced Explanation Instability in Graph Neural Networks","display_name":"SOFIA: Quantifying Sampling-Induced Explanation Instability in Graph Neural Networks","publication_year":2026,"publication_date":"2026-11-11","ids":{"openalex":"https://openalex.org/W7225808538","doi":"https://doi.org/10.21203/rs.3.rs-11136263/v1"},"language":"en","primary_location":{"id":"doi:10.21203/rs.3.rs-11136263/v1","doi":"https://doi.org/10.21203/rs.3.rs-11136263/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11136263/v1","pdf_url":null,"source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"Research Square","raw_type":"posted-content"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.21203/rs.3.rs-11136263/v1","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5161439858","display_name":"Prabu N P","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I145286018","display_name":"SRM Institute of Science and Technology","ror":"https://ror.org/050113w36","country_code":"IN","type":"education","lineage":["https://openalex.org/I145286018"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Prabu N P","raw_affiliation_strings":["SRM Institute of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"SRM Institute of Science and Technology","institution_ids":["https://openalex.org/I145286018"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039154929","display_name":"Priyanga Subbiah","orcid":"https://orcid.org/0000-0002-2395-7492","observed_orcids":["https://orcid.org/0000-0002-2395-7492"]},"institutions":[{"id":"https://openalex.org/I145286018","display_name":"SRM Institute of Science and Technology","ror":"https://ror.org/050113w36","country_code":"IN","type":"education","lineage":["https://openalex.org/I145286018"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Priyanga Subbiah","raw_affiliation_strings":["SRM Institute of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"SRM Institute of Science and Technology","institution_ids":["https://openalex.org/I145286018"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5161481525","display_name":"Daniel A","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I164861460","display_name":"Manipal Academy of Higher Education","ror":"https://ror.org/02xzytt36","country_code":"IN","type":"education","lineage":["https://openalex.org/I164861460"]}],"countries":["IN"],"is_corresponding":true,"raw_author_name":"Daniel A","raw_affiliation_strings":["Manipal Academy of Higher Education"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Manipal Academy of Higher Education","institution_ids":["https://openalex.org/I164861460"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5161481525"],"corresponding_institution_ids":["https://openalex.org/I164861460"],"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.9990000128746033,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.9990000128746033,"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.0008999999845400453,"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/T10064","display_name":"Complex Network Analysis Techniques","score":0.0,"subfield":{"id":"https://openalex.org/subfields/3109","display_name":"Statistical and Nonlinear Physics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[],"concepts":[{"id":"https://openalex.org/C2779812341","wikidata":"https://www.wikidata.org/wiki/Q5891525","display_name":"Homophily","level":2,"score":0.6044999957084656},{"id":"https://openalex.org/C203519979","wikidata":"https://www.wikidata.org/wiki/Q865360","display_name":"Jaccard index","level":3,"score":0.5392000079154968},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.5320000052452087},{"id":"https://openalex.org/C207821765","wikidata":"https://www.wikidata.org/wiki/Q405372","display_name":"Instability","level":2,"score":0.521399974822998},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.49720001220703125},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4577000141143799},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.4097999930381775},{"id":"https://openalex.org/C173362246","wikidata":"https://www.wikidata.org/wiki/Q8216024","display_name":"Visibility graph","level":3,"score":0.39719998836517334},{"id":"https://openalex.org/C149629883","wikidata":"https://www.wikidata.org/wiki/Q660926","display_name":"Fraction (chemistry)","level":2,"score":0.38960000872612},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.37540000677108765},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.36169999837875366},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.34049999713897705},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.3328000009059906},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.32510000467300415},{"id":"https://openalex.org/C47385372","wikidata":"https://www.wikidata.org/wiki/Q7098943","display_name":"Optimal distinctiveness theory","level":2,"score":0.32179999351501465},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.3199000060558319},{"id":"https://openalex.org/C2775997480","wikidata":"https://www.wikidata.org/wiki/Q586277","display_name":"Degree (music)","level":2,"score":0.31929999589920044},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.31029999256134033},{"id":"https://openalex.org/C66882249","wikidata":"https://www.wikidata.org/wiki/Q169336","display_name":"Homogeneous","level":2,"score":0.3075999915599823},{"id":"https://openalex.org/C62611344","wikidata":"https://www.wikidata.org/wiki/Q1062658","display_name":"Node (physics)","level":2,"score":0.3043000102043152},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.2872999906539917},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.2824000120162964},{"id":"https://openalex.org/C64708745","wikidata":"https://www.wikidata.org/wiki/Q2998010","display_name":"Partial correlation","level":3,"score":0.2687000036239624},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.2671999931335449},{"id":"https://openalex.org/C117220453","wikidata":"https://www.wikidata.org/wiki/Q5172842","display_name":"Correlation","level":2,"score":0.26649999618530273},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.262800008058548},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.26269999146461487},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.2583000063896179},{"id":"https://openalex.org/C2777035058","wikidata":"https://www.wikidata.org/wiki/Q1662634","display_name":"Clique","level":2,"score":0.2515999972820282},{"id":"https://openalex.org/C88230418","wikidata":"https://www.wikidata.org/wiki/Q131476","display_name":"Graph theory","level":2,"score":0.25029999017715454}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21203/rs.3.rs-11136263/v1","doi":"https://doi.org/10.21203/rs.3.rs-11136263/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11136263/v1","pdf_url":null,"source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"Research Square","raw_type":"posted-content"}],"best_oa_location":{"id":"doi:10.21203/rs.3.rs-11136263/v1","doi":"https://doi.org/10.21203/rs.3.rs-11136263/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11136263/v1","pdf_url":null,"source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"Research Square","raw_type":"posted-content"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":19,"referenced_works":["https://openalex.org/W2153959628","https://openalex.org/W2786915849","https://openalex.org/W2918342466","https://openalex.org/W2961295589","https://openalex.org/W2975966238","https://openalex.org/W2995509042","https://openalex.org/W3091534907","https://openalex.org/W3108823960","https://openalex.org/W3116637551","https://openalex.org/W3208238874","https://openalex.org/W4290927864","https://openalex.org/W4321853830","https://openalex.org/W4327813569","https://openalex.org/W4387008453","https://openalex.org/W4387355946","https://openalex.org/W7133192667","https://openalex.org/W7133211418","https://openalex.org/W7133214608","https://openalex.org/W7133227520"],"related_works":[],"abstract_inverted_index":{"Abstract":[0],"Post-hoc":[1],"explainability":[2,69],"tools":[3],"for":[4,66,132,310],"Graph":[5,97],"Neural":[6],"Networks":[7,99],"rest":[8],"on":[9,100],"an":[10,125,290],"implicit":[11],"assumption:":[12],"that":[13,48,240,271,279],"the":[14,27,34,75,155,188,238,272,303],"explanations":[15,120,282],"they":[16],"produce":[17],"reflect":[18],"stable":[19],"properties":[20],"of":[21,80,93,170,172,185,224,249],"a":[22,43,55,64,161,168,182,211,214,219,255,306],"node's":[23,51],"structural":[24,298],"role":[25],"in":[26,71,243,283],"graph.":[28],"This":[29],"assumption":[30],"is":[31,274],"violated":[32],"whenever":[33],"underlying":[35],"model":[36],"was":[37,262],"trained":[38],"with":[39,147,187,237],"neighbor":[40,83],"sampling":[41,84,166],"\u2014":[42,111,116],"universally":[44],"adopted":[45],"scalability":[46],"practice":[47],"restricts":[49],"each":[50],"visible":[52],"neighborhood":[53,152],"to":[54,179],"randomly":[56],"drawn":[57],"subset":[58],"during":[59],"training.":[60],"We":[61],"introduce":[62],"Sofia,":[63],"framework":[65],"measuring":[67],"Sampling-induced":[68],"instability":[70,195,247,273],"GNNs,":[72],"and":[73,95,108,114,117,136,301],"present":[74],"first":[76],"systematic":[77],"empirical":[78],"study":[79],"how":[81],"training-time":[82,151],"affects":[85],"post-hoc":[86,280],"explanation":[87,176],"outputs.":[88],"Sofia":[89],"trains":[90],"matched":[91],"pairs":[92],"full-neighborhood":[94,148],"sampled":[96,142],"Convolutional":[98],"three":[101],"real-world":[102],"datasets":[103],"spanning":[104],"different":[105],"homophily":[106],"levels":[107],"degree":[109,197],"distributions":[110],"Cora,":[112,254],"Reddit,":[113,160],"Actor":[115],"compares":[118],"their":[119],"using":[121],"Jaccard":[122],"overlap":[123],"against":[124,181],"empirically":[126],"measured":[127],"same-model":[128],"noise":[129,183],"floor,":[130],"controlling":[131],"both":[133],"explainer":[134],"stochasticity":[135],"node":[137,299],"degree.":[138],"Across":[139],"all":[140],"conditions,":[141],"models":[143],"maintain":[144],"accuracy":[145],"parity":[146],"models,":[149],"isolating":[150],"visibility":[153],"as":[154,305],"sole":[156],"varying":[157],"factor.":[158],"On":[159,209,253],"high-degree":[162],"graph":[163],"where":[164],"aggressive":[165],"excluded":[167],"mean":[169,221],"89.7%":[171],"true":[173],"neighbors,":[174],"cross-model":[175],"agreement":[177],"dropped":[178],"0.292":[180],"floor":[184],"0.538,":[186],"influential-neighbor":[189],"exclusion":[190,222,251],"fraction":[191,223],"significantly":[192],"predicting":[193],"per-node":[194,308],"after":[196],"control":[198],"(\u03b2":[199,226,264],"=":[200,203,207,227,234,265,268],"0.687,":[201],"p":[202,229,267],"3.33e-15,":[204],"partial":[205,232],"R\u00b2":[206,233],"0.095).":[208],"Actor,":[210],"heterophilous":[212,244],"graph,":[213,258],"significant":[215,260],"effect":[216,261],"emerged":[217],"at":[218],"lower":[220],"71.1%":[225],"0.255,":[228],"<":[230],"0.001,":[231],"0.130),":[235],"consistent":[236],"prediction":[239],"signal":[241],"concentration":[242],"neighborhoods":[245],"amplifies":[246],"independently":[248],"raw":[250],"magnitude.":[252],"low-degree":[256],"homophilous":[257],"no":[259],"observed":[263],"-0.116,":[266],"0.401),":[269],"confirming":[270],"structure-dependent.":[275],"These":[276],"results":[277],"demonstrate":[278],"GNN":[281],"production":[284],"systems":[285],"are":[286],"partially":[287],"determined":[288],"by":[289,296],"arbitrary":[291],"training":[292],"hyperparameter":[293],"rather":[294],"than":[295],"genuine":[297],"properties,":[300],"establish":[302],"SOFIA-Index":[304],"practical":[307],"diagnostic":[309],"sampling-contaminated":[311],"explanations.":[312]},"counts_by_year":[],"updated_date":"2026-11-12T12:28:57.059296","created_date":"2026-11-12T00:00:00"}