{"id":"W3175558129","doi":"10.1609/aaai.v35i12.17276","title":"Improving Sample Efficiency in Model-Free Reinforcement Learning from Images","year":2021,"lang":"en","type":"article","venue":"","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":201,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Reinforcement learning; Robustness (evolution); Computer science; Artificial intelligence; Stability (learning theory); Machine learning; Encoder; Representation (politics); Noise (video); Code (set theory); Image (mathematics)","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":[],"consensus_categories":[],"category_scores_codex":[0.0003401872,0.0001889616,0.0002121069,0.0001290999,0.0001490738,0.0003498227,0.001193542,0.00007138411,0.0001303387],"category_scores_gemma":[0.001040022,0.0001903449,0.00007206883,0.0005173096,0.00003073042,0.0006324667,0.001483019,0.0003685573,0.00004491041],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001265539,"about_ca_system_score_gemma":0.0002101923,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007857587,"about_ca_topic_score_gemma":0.00003856934,"domain_scores_codex":[0.9979848,0.0000757864,0.0004448449,0.0005542304,0.0004433271,0.0004969796],"domain_scores_gemma":[0.9982879,0.0003288337,0.0001328105,0.001045248,0.0001121638,0.00009301617],"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":[0.000001690143,0.00001546216,0.002173952,0.000008029428,0.000005836173,0.00001594585,0.0003893141,0.9806615,0.002360051,0.01152122,0.0001284205,0.002718571],"study_design_scores_gemma":[0.0005247236,0.00004483081,0.0005192518,0.00002398881,0.000003737429,0.000001846428,0.00007204143,0.9921091,0.004825045,0.001455438,0.0001816641,0.0002383008],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002079017,0.00008132191,0.9859313,0.0004868007,0.0001821406,0.0001074023,6.640364e-7,0.0002480983,0.01088325],"genre_scores_gemma":[0.6202108,0.00002670959,0.3755634,0.0003490996,0.00003364306,0.00001006844,0.00001966113,0.00001538164,0.003771283],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6181317,"threshold_uncertainty_score":0.7762042,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01918541106757507,"score_gpt":0.2364670789210525,"score_spread":0.2172816678534774,"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."}}