{"id":"W3151722529","doi":"10.15607/rss.2021.xvii.012","title":"Learning Generalizable Robotic Reward Functions from “In-The-Wild” Human Videos","year":2021,"lang":"en","type":"preprint","venue":"","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Office of Naval Research; Canadian Institute for Advanced Research; National Science Foundation","keywords":"Computer science; Artificial intelligence; Generalization; Reinforcement learning; Task (project management); Robot; Function (biology); Discriminator; Machine learning; Robotics; Human–computer interaction","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001088039,0.001380503,0.0008637428,0.0004918876,0.0002326904,0.0006180445,0.001361483,0.001495935,0.001277924],"category_scores_gemma":[0.005071912,0.0004330018,0.0006814766,0.0003429175,0.001111862,0.001665567,0.00105217,0.001946753,0.0003636781],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001059197,"about_ca_system_score_gemma":0.0006863284,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004224854,"about_ca_topic_score_gemma":0.005683206,"domain_scores_codex":[0.9995003,0.0001213307,0.00001655799,0.0002420176,0.00005715154,0.00006261408],"domain_scores_gemma":[0.998743,0.0006492925,0.0001811099,0.0001823004,0.0001177336,0.0001265779],"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":[0.000498554,0.0004508068,0.009167006,0.0002195801,0.0001589059,0.0003016405,0.0001577626,0.7710474,0.01323472,0.007332157,0.006144728,0.1912869],"study_design_scores_gemma":[0.00002029386,0.0001023352,0.001289431,0.00001334137,0.000008435757,0.00004742582,0.00001707133,0.9884279,0.001881205,0.007684625,0.0004944477,0.00001341412],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.241898,0.001004168,0.7488403,0.00127468,0.0001210907,0.0001845678,0.001192656,0.002638788,0.002845807],"genre_scores_gemma":[0.9191018,0.0003186976,0.07485108,0.0005742189,0.00007875833,0.0001556185,0.001763869,0.000120389,0.003035511],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004224854,"threshold_uncertainty_score":0.008400559,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04143465606680853,"score_gpt":0.2716297009467223,"score_spread":0.2301950448799138,"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."}}