{"id":"W2977481643","doi":"10.1609/aaai.v35i12.17276","title":"Improving Sample Efficiency in Model-Free Reinforcement Learning from Images","year":2021,"lang":"en","type":"preprint","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":121,"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":["metaepi_narrow","scholarly_communication","open_science"],"consensus_categories":[],"category_scores_codex":[0.001004913,0.0006026833,0.0007198311,0.0003464766,0.0002590519,0.001234985,0.006937353,0.0003549366,0.00008003823],"category_scores_gemma":[0.003694564,0.0005357641,0.0002971207,0.0007600604,0.0002785183,0.0005825539,0.007959411,0.002202729,0.0000212157],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002768853,"about_ca_system_score_gemma":0.0006024126,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001368372,"about_ca_topic_score_gemma":0.00003726528,"domain_scores_codex":[0.9950778,0.0000568407,0.001467782,0.001345678,0.001311992,0.0007399079],"domain_scores_gemma":[0.9958242,0.0003456781,0.001469649,0.001260296,0.0009707852,0.0001293853],"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":[0.00002481621,0.0000725728,0.0002966763,0.0001345743,0.00002212673,0.000001539708,0.00277178,0.9024622,0.008960551,0.07695018,0.00002363124,0.008279298],"study_design_scores_gemma":[0.00004771079,0.0001026249,0.00005280404,0.0009298519,0.0000202979,6.780675e-7,0.0005437651,0.840656,0.1144794,0.04271866,0.000004452451,0.000443702],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02660413,0.00006107363,0.9663233,0.001035513,0.0007650839,0.000698654,0.000006165523,0.0001626365,0.004343485],"genre_scores_gemma":[0.9547352,0.0001270102,0.04452452,0.0001286741,0.00007668362,0.00006649292,0.0000101102,0.00003510881,0.0002961514],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9281311,"threshold_uncertainty_score":0.9998018,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07127289501225326,"score_gpt":0.2836056152858242,"score_spread":0.2123327202735709,"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."}}