{"id":"W2970987521","doi":"10.1007/978-3-030-26250-1_23","title":"Improving ML Safety with Partial Specifications","year":2019,"lang":"en","type":"book-chapter","venue":"Lecture notes in computer science","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Workflow; Software engineering; Key (lock); Component (thermodynamics); Safety assurance; Software; Partial evaluation; Programming language; Reliability engineering; Operating system; Database; Engineering","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"],"consensus_categories":[],"category_scores_codex":[0.0009104038,0.0005789486,0.0005577253,0.0006420342,0.0004331282,0.0006251687,0.004079711,0.0002940495,0.00006406964],"category_scores_gemma":[0.0001355864,0.0004954074,0.0001093603,0.0007110527,0.0006781255,0.0009219946,0.001656085,0.00142531,0.0001436984],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004427087,"about_ca_system_score_gemma":0.001230155,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003775286,"about_ca_topic_score_gemma":0.00005673951,"domain_scores_codex":[0.9955124,0.00005206523,0.0005448174,0.001955523,0.001172526,0.0007626684],"domain_scores_gemma":[0.9962944,0.0005899746,0.0004847435,0.002212658,0.000245326,0.0001728825],"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":[0.00001637209,0.00001732729,0.0001686542,0.00002981058,0.00001251469,0.00006015856,0.0004942925,0.4367118,0.00007682171,0.0895634,0.000006923814,0.4728419],"study_design_scores_gemma":[0.0004445497,0.0001891725,0.0002165935,0.0002726033,0.00001395927,0.00008338757,2.935944e-7,0.9788212,0.0003455858,0.01315747,0.005577905,0.0008772874],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00001943375,0.0001063761,0.9878809,0.0009350231,0.001871651,0.0004910274,0.000003825513,0.0002532203,0.008438489],"genre_scores_gemma":[0.08550688,0.00001580751,0.911414,0.0007028072,0.0009868458,0.000007820669,0.00000987552,0.00007223174,0.001283689],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.5421094,"threshold_uncertainty_score":0.9997498,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01840434861093901,"score_gpt":0.2362812111301523,"score_spread":0.2178768625192133,"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."}}