{"id":"W4417267046","doi":"10.1007/s10618-025-01180-w","title":"Certifying robustness of graph convolutional networks for node perturbation with polyhedra abstract interpretation","year":2025,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"McGill University","funders":"Fonds de recherche du Québec – Nature et technologies; Nemzeti Kutatási Fejlesztési és Innovációs Hivatal; Mitacs; Knut och Alice Wallenbergs Stiftelse","keywords":"Robustness (evolution); Graph; Convolutional neural network; Adversarial system; Polyhedron; Training set","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001962275,0.0001371937,0.0001807224,0.0001407043,0.0001441112,0.000130402,0.0005319394,0.0000595924,4.951567e-7],"category_scores_gemma":[0.00006867203,0.0001162421,0.00003233689,0.000380948,0.0001040753,0.001760836,0.0002929944,0.00008693375,1.094855e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000158467,"about_ca_system_score_gemma":0.00008272872,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004785639,"about_ca_topic_score_gemma":0.00004242464,"domain_scores_codex":[0.9990012,0.00002623866,0.0002454489,0.0004531824,0.00008372837,0.0001902079],"domain_scores_gemma":[0.9987323,0.0004977957,0.0001377869,0.0004965035,0.0001031825,0.00003240101],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00169849,0.0005271618,0.01786839,0.001076374,0.0005186288,0.000006633192,0.002997268,0.2440348,0.000728432,0.07295075,0.009889998,0.6477031],"study_design_scores_gemma":[0.00046127,0.00004961113,0.007016617,0.0004271077,0.00002893058,0.00000381256,0.0001504233,0.9910749,0.00006735479,0.0004479843,0.0001246528,0.0001473357],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03744911,0.001630023,0.9599305,0.00009015579,0.0004375163,0.0001585167,0.00006549373,0.00004080784,0.0001978542],"genre_scores_gemma":[0.965618,0.00004863987,0.03360411,0.0000483242,0.00006486536,0.00002077352,0.000416216,0.000008043251,0.0001710196],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9281689,"threshold_uncertainty_score":0.4740216,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03197076562914049,"score_gpt":0.2958814579966345,"score_spread":0.263910692367494,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}