{"id":"W4391725330","doi":"10.14722/ndss.2024.23014","title":"Overconfidence is a Dangerous Thing: Mitigating Membership Inference Attacks by Enforcing Less Confident Prediction","year":2024,"lang":"en","type":"article","venue":"","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":21,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Overconfidence effect; Inference; Computer security; Computer science; Artificial intelligence; Psychology; Social psychology","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008794163,0.001107594,0.001299427,0.0007565381,0.001587821,0.002520223,0.002717914,0.002326096,0.0009958968],"category_scores_gemma":[0.03710253,0.0006237502,0.001014494,0.0008453091,0.002432389,0.006354056,0.00682379,0.005949405,0.0005921081],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001322879,"about_ca_system_score_gemma":0.001895275,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009945466,"about_ca_topic_score_gemma":0.001025323,"domain_scores_codex":[0.9865611,0.005505663,0.0006342944,0.00200515,0.004381567,0.0009122535],"domain_scores_gemma":[0.9684991,0.01196546,0.003330441,0.01387371,0.001653997,0.0006773762],"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":[0.001808093,0.0007065324,0.02899543,0.000383278,0.0004216171,0.0009304072,0.001855035,0.2914284,0.03836417,0.1190051,0.0158547,0.5002472],"study_design_scores_gemma":[0.00007215256,0.0002734632,0.002430056,0.00006493356,0.00007125275,0.0006736955,0.0001764424,0.9196941,0.026266,0.04509178,0.005120579,0.00006550778],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.183435,0.001185252,0.8019205,0.00376569,0.0001658295,0.0001821018,0.0002555517,0.002992886,0.006097262],"genre_scores_gemma":[0.9474428,0.0001896996,0.05028046,0.000596726,0.0001173788,0.00006683596,0.0001572784,0.00007764281,0.001071292],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008794163,"threshold_uncertainty_score":0.04650849,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2179136209609521,"score_gpt":0.4373605369772627,"score_spread":0.2194469160163106,"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."}}