{"id":"W4416048252","doi":"10.48550/arxiv.2505.22356","title":"Suitability Filter: A Statistical Framework for Classifier Evaluation in Real-World Deployment Settings","year":2025,"lang":"en","type":"preprint","venue":"ArXiv.org","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Government of Canada; Canadian Institute for Advanced Research","keywords":"Classifier (UML); Ground truth; Covariate; Modular design; Filter (signal processing); Software deployment; Statistical hypothesis testing; Test data; Statistical model","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.003623284,0.0004236982,0.0006055383,0.0003281738,0.0001748618,0.0001774908,0.00151466,0.0004330132,0.00009196971],"category_scores_gemma":[0.004958029,0.0004490154,0.0001668863,0.0005275373,0.00009382557,0.0002255612,0.002462332,0.001786907,0.00002069955],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001001949,"about_ca_system_score_gemma":0.0008576834,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005390542,"about_ca_topic_score_gemma":0.0004623066,"domain_scores_codex":[0.9954333,0.0007877562,0.0008583359,0.001576106,0.0007261666,0.0006183675],"domain_scores_gemma":[0.9945807,0.002976192,0.0004044084,0.001599918,0.0003174342,0.0001213164],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0000840515,0.0002249196,0.777355,0.000602795,0.00008858044,0.00001385343,0.001468442,0.05080648,0.00001140559,0.1221215,0.001270835,0.04595216],"study_design_scores_gemma":[0.0005143347,0.00003958307,0.3832158,0.0004917016,0.0000677131,4.438702e-7,0.00003003643,0.4856515,0.00005496035,0.1284914,0.001027122,0.0004153996],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07628836,0.00004743521,0.9153882,0.003317409,0.001919786,0.00177698,0.0000521574,0.0002533679,0.000956293],"genre_scores_gemma":[0.5975775,0.00001161819,0.4004042,0.0005085711,0.0002573368,0.0008027134,0.0001083525,0.00002745443,0.0003022773],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.5212892,"threshold_uncertainty_score":0.9997962,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08535396798824715,"score_gpt":0.3860397162206754,"score_spread":0.3006857482324282,"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."}}