{"id":"W4380089110","doi":"10.2139/ssrn.4474509","title":"Calibration Attack: Adversarial Attacks Against Model Calibration","year":2023,"lang":"en","type":"preprint","venue":"SSRN Electronic Journal","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"ca_institutions":"National Research Council Canada; Queen's University","funders":"","keywords":"Adversarial system; Calibration; Computer science; Computer security; Artificial intelligence; Mathematics; Statistics","routes":{"ca_aff":true,"ca_fund":false,"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":["metaepi_narrow","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.003466673,0.0006464071,0.0006125005,0.0005341037,0.0006909494,0.0009705464,0.002905522,0.0007407626,0.000007923712],"category_scores_gemma":[0.0003844769,0.0006670064,0.000443198,0.000518764,0.00007445018,0.001641008,0.002181677,0.01027366,0.00004964104],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002557912,"about_ca_system_score_gemma":0.01070119,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009229057,"about_ca_topic_score_gemma":0.0004138518,"domain_scores_codex":[0.9929951,0.0005467481,0.0009933264,0.001142449,0.001173058,0.003149271],"domain_scores_gemma":[0.9972815,0.0001581574,0.001038834,0.001087794,0.0002098214,0.0002239214],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00003156418,0.00003201074,0.000123729,0.00002211305,0.0001726412,0.00001796284,0.000437619,0.8924916,0.00005717774,0.1003663,0.0007761151,0.005471186],"study_design_scores_gemma":[0.0005648785,0.00006945883,0.0000178459,0.00007438617,0.00004334427,0.00004855303,0.0001085391,0.7655513,0.00002229276,0.2328724,0.0001242825,0.0005026691],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.006734232,0.0003180578,0.9835593,0.003937431,0.003872665,0.0004190837,0.00001021895,0.0007449538,0.000404117],"genre_scores_gemma":[0.952764,0.002769512,0.03389473,0.0005328039,0.004462805,0.00006134689,0.0002754504,0.000253672,0.004985727],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9496645,"threshold_uncertainty_score":0.9995781,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03142698594566107,"score_gpt":0.2950429123845449,"score_spread":0.2636159264388838,"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."}}