{"id":"W4327809167","doi":"10.1109/tsmc.2023.3250120","title":"Spatial Hard Attention Modeling via Deep Reinforcement Learning for Skeleton-Based Human Activity Recognition","year":2023,"lang":"en","type":"article","venue":"IEEE Transactions on Systems Man and Cybernetics Systems","topic":"Human Pose and Action Recognition","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Discriminative model; Artificial intelligence; Computer science; Skeleton (computer programming); Reinforcement learning; Deep learning; Activity recognition; Pattern recognition (psychology); Human skeleton; Hidden Markov model; Machine learning; Frame (networking); Process (computing); Joint (building); 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":[0.0005951113,0.0002876749,0.000335329,0.0004816887,0.0008331339,0.0004978796,0.0001926315,0.0001902495,0.000007628751],"category_scores_gemma":[0.000004428784,0.0003073201,0.0001718042,0.0003228881,0.0000294706,0.0003512653,0.000004087218,0.0002760846,0.0001489765],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001391464,"about_ca_system_score_gemma":0.00003196294,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008143338,"about_ca_topic_score_gemma":0.00008736086,"domain_scores_codex":[0.997757,0.0002260218,0.0005550492,0.0005955779,0.0004555223,0.0004108922],"domain_scores_gemma":[0.9989117,0.000117611,0.000258419,0.0003252412,0.0002322912,0.0001547298],"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":[0.00004302377,0.0001044958,0.00001609718,0.000525729,0.0000884097,0.000004954042,0.0002994871,0.946593,0.008987157,0.0001556172,0.00006346792,0.0431186],"study_design_scores_gemma":[0.001015638,0.0004009021,0.00005974834,0.000316723,0.00005235147,0.00001492648,0.0001789795,0.9955417,0.001699084,0.00008040394,0.000283153,0.0003564135],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1107973,0.00002198702,0.8855933,0.00005025679,0.001701866,0.001074631,0.00000870543,0.0005230819,0.0002289069],"genre_scores_gemma":[0.9973652,0.00002763451,0.0001495129,0.00002448704,0.000219929,0.0005923116,0.00006108904,0.00003971172,0.001520108],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.886568,"threshold_uncertainty_score":0.9999379,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04223951700842035,"score_gpt":0.2555025766333077,"score_spread":0.2132630596248874,"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."}}