{"id":"W4281553483","doi":"10.24963/ijcai.2022/484","title":"Approximate Exploitability: Learning a Best Response","year":2022,"lang":"en","type":"article","venue":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","topic":"Artificial Intelligence in Games","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Adversarial system; Computer science; Reinforcement learning; Scalability; Robustness (evolution); Artificial intelligence; Machine learning; Variety (cybernetics); Computation; Deep neural networks; Artificial neural network; Theoretical computer science; Algorithm","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":[0.002789881,0.0004025043,0.0004234739,0.0004554369,0.001081287,0.0005190195,0.00536157,0.0000935223,0.0006589162],"category_scores_gemma":[0.002362454,0.0003497649,0.0003387607,0.00107293,0.0004934043,0.0008332912,0.002895761,0.001165012,0.0002080748],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000437118,"about_ca_system_score_gemma":0.0002246251,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001000764,"about_ca_topic_score_gemma":0.00001530121,"domain_scores_codex":[0.9950964,0.0001609882,0.001271506,0.0009944926,0.001897285,0.0005793254],"domain_scores_gemma":[0.9966922,0.0004440254,0.0009454505,0.0005990957,0.001171153,0.0001480913],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0007377137,0.0007059451,0.0007355756,0.00003032374,0.00005684322,0.000005430215,0.006125017,0.005940164,0.03582571,0.921765,0.0002178166,0.02785442],"study_design_scores_gemma":[0.00005973458,0.0009463301,0.0002323905,0.0002710725,0.0000189351,0.00005418564,0.01104608,0.2818469,0.4018309,0.2976894,0.005342788,0.0006613152],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7861026,0.00006615539,0.1085731,0.06501894,0.005460396,0.002688464,0.00006224321,0.0008319143,0.03119626],"genre_scores_gemma":[0.9944797,0.00002828804,0.003545207,0.0003784286,0.0001172698,0.0003961323,0.000001763487,0.00003268664,0.00102058],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6240757,"threshold_uncertainty_score":0.9998955,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08809639599173266,"score_gpt":0.2973602486952115,"score_spread":0.2092638527034789,"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."}}