{"id":"W4377971347","doi":"10.1109/tai.2023.3279057","title":"Interpreting Tangled Program Graphs Under Partially Observable Dota 2 Invoker Tasks","year":2023,"lang":"en","type":"article","venue":"IEEE Transactions on Artificial Intelligence","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Computer science; Interpretability; Artificial intelligence; Graph; Context (archaeology); Machine learning; Task (project management); Theoretical computer science","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","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0005934192,0.0003170495,0.0002620575,0.0004118471,0.0005176862,0.0005474429,0.001188712,0.0001574909,0.0001527936],"category_scores_gemma":[0.00004629457,0.0003218475,0.0002225065,0.002382001,0.0001749651,0.0006121966,0.00002046531,0.0005811628,0.002045549],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001029142,"about_ca_system_score_gemma":0.0001100824,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007527359,"about_ca_topic_score_gemma":0.0001331296,"domain_scores_codex":[0.9970394,0.0001520404,0.0007217196,0.000702921,0.0006191909,0.0007647194],"domain_scores_gemma":[0.9982167,0.0003596954,0.0001605773,0.000892935,0.0001729813,0.0001970431],"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":[0.00002245999,0.0001330918,0.0000103954,0.00001307189,0.00004129376,0.00001738742,0.0007060345,0.8544594,0.001848581,0.01554655,0.0001508995,0.1270508],"study_design_scores_gemma":[0.00003876991,0.0003404067,0.00002976855,0.00008152886,0.00001859228,0.000005525159,0.0002881334,0.9116938,0.07731156,0.008863362,0.0009610455,0.0003675241],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.005704484,0.000008894842,0.9876691,0.001157325,0.002149243,0.0005689865,0.000003443661,0.001873617,0.0008649002],"genre_scores_gemma":[0.9826013,0.00004459998,0.01562555,0.0005101623,0.00004697152,0.0001881371,0.000004321351,0.00004012918,0.0009388016],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9768968,"threshold_uncertainty_score":0.9999233,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08019492026766367,"score_gpt":0.3238346140799372,"score_spread":0.2436396938122736,"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."}}