{"id":"W3116666954","doi":"10.1109/spw50608.2020.00027","title":"On the Robustness of Cooperative Multi-Agent Reinforcement Learning","year":2020,"lang":"en","type":"article","venue":"","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":56,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute; University of Toronto","funders":"Canadian Institute for Advanced Research","keywords":"Reinforcement learning; Adversary; Robustness (evolution); Adversarial system; Computer science; Harm; Artificial intelligence; Computer security; Benchmark (surveying); Differentiable function; Mathematics; Psychology; Social psychology","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":[0.005973188,0.001330499,0.001020918,0.0008346715,0.0005944173,0.0009065717,0.001208783,0.001164062,0.001193207],"category_scores_gemma":[0.03023584,0.0004119898,0.0006486656,0.0003250259,0.002178628,0.001484636,0.00233676,0.002151326,0.0002637169],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00144027,"about_ca_system_score_gemma":0.001209343,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004205626,"about_ca_topic_score_gemma":0.001848408,"domain_scores_codex":[0.9975466,0.001040856,0.0001152116,0.0004266687,0.0005027931,0.0003678527],"domain_scores_gemma":[0.9771516,0.01741887,0.001904107,0.001609634,0.001192909,0.0007229107],"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":[0.0001137491,0.00004373136,0.001132974,0.00003768909,0.00004934211,0.00004039317,0.00003613562,0.9859962,0.0009403954,0.004577403,0.0002991747,0.006732863],"study_design_scores_gemma":[0.000007694341,0.0000642851,0.0001585957,0.000007250906,0.000005569222,0.00001026173,0.000007181654,0.9956132,0.0003807601,0.003643887,0.00009667506,0.000004630761],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2833044,0.001475038,0.7040144,0.001361068,0.0001446985,0.0001430132,0.0001465742,0.001024135,0.00838667],"genre_scores_gemma":[0.9848427,0.000154455,0.01398385,0.000117435,0.00002593882,0.00005008891,0.00005762488,0.00004821568,0.0007198431],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005973188,"threshold_uncertainty_score":0.03158963,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04048919975561101,"score_gpt":0.27154417538606,"score_spread":0.231054975630449,"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."}}