{"id":"W4402264355","doi":"10.1109/sp54263.2024.00243","title":"SoK: Unintended Interactions among Machine Learning Defenses and Risks","year":2024,"lang":"en","type":"article","venue":"","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"Government of Ontario","keywords":"Unintended consequences; Computer science; Machine learning; Human–computer interaction; Risk analysis (engineering); Business; Political science","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":[0.009419977,0.001156511,0.001426022,0.001154716,0.001841719,0.004247387,0.00224716,0.003651204,0.00679525],"category_scores_gemma":[0.06639133,0.0007472599,0.00115369,0.0006855747,0.008596735,0.008901156,0.005442859,0.00685957,0.0006704614],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002211735,"about_ca_system_score_gemma":0.002049026,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007209274,"about_ca_topic_score_gemma":0.000728277,"domain_scores_codex":[0.9916128,0.003782892,0.0003315487,0.001510624,0.001725455,0.001036553],"domain_scores_gemma":[0.9282599,0.04804098,0.007764877,0.01169188,0.002571794,0.001670513],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001730453,0.00007480491,0.003753163,0.0001610984,0.0001241729,0.0002445569,0.0004589599,0.06736954,0.00192196,0.8966521,0.004551545,0.02451499],"study_design_scores_gemma":[0.00001854299,0.00006627152,0.0008270777,0.00006055309,0.00003383053,0.0002209942,0.0001115965,0.1188941,0.0008479325,0.8766449,0.002230625,0.00004357784],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.07871021,0.001180763,0.8687602,0.01219966,0.0003437361,0.0001694982,0.0002770166,0.0007919476,0.03756689],"genre_scores_gemma":[0.9615389,0.0004284068,0.03216459,0.001318883,0.0002211356,0.0001593908,0.00007762163,0.0001570553,0.003934025],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.009419977,"threshold_uncertainty_score":0.04981822,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02972554940440073,"score_gpt":0.3047028500296028,"score_spread":0.2749773006252021,"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."}}