{"id":"W3212787291","doi":"10.1109/qrs54544.2021.00037","title":"On Assessing The Safety of Reinforcement Learning algorithms Using Formal Methods","year":2021,"lang":"en","type":"article","venue":"2021 IEEE 21st International Conference on Software Quality, Reliability and Security (QRS)","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Adversarial system; Probabilistic logic; Reinforcement learning; Computer science; Artificial intelligence; Computer security; Multi-agent system; Autonomous agent; Grid; Machine learning; Mathematics","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":[0.005035868,0.0003166614,0.0004874633,0.0001170844,0.000608278,0.0004703234,0.001022099,0.0001884002,0.0002918755],"category_scores_gemma":[0.004493813,0.0002696184,0.0002091481,0.0003732582,0.0002753265,0.0009351127,0.0008479935,0.001178899,0.000006725729],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000232891,"about_ca_system_score_gemma":0.0004102301,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002480139,"about_ca_topic_score_gemma":0.00001491745,"domain_scores_codex":[0.9946704,0.002061088,0.0009475261,0.0008202532,0.001107306,0.0003934379],"domain_scores_gemma":[0.9950376,0.002498257,0.0005769003,0.0007907971,0.0009737628,0.0001226731],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001100543,0.0002679261,0.003343037,0.0001585825,0.0001416347,0.00001723995,0.004663981,0.1742697,0.0005447029,0.7381063,0.00001614069,0.07836075],"study_design_scores_gemma":[0.0007923249,0.0001664227,0.002337439,0.0003276605,0.00002836497,0.0000244783,0.001602941,0.9389766,0.002543772,0.05172563,0.001015042,0.0004593172],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07013962,0.00003209411,0.9236363,0.001443794,0.001347638,0.000191777,0.00001381267,0.00008018511,0.003114766],"genre_scores_gemma":[0.8748657,0.0000911314,0.1242753,0.0003943564,0.0001531027,0.00001042349,0.00003225791,0.0000148568,0.0001628838],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8047261,"threshold_uncertainty_score":0.9999756,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06773412524589584,"score_gpt":0.4105192365666844,"score_spread":0.3427851113207886,"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."}}