{"id":"W4415021619","doi":"10.1016/j.arcontrol.2025.101027","title":"A view on learning robust goal-conditioned value functions: Interplay between RL and MPC","year":2025,"lang":"en","type":"article","venue":"Annual Reviews in Control","topic":"Advanced Control Systems Optimization","field":"Engineering","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"Honeywell (Canada); University of British Columbia","funders":"Fusion Energy Sciences; Office of Science; U.S. Department of Energy","keywords":"Robustness (evolution); Reinforcement learning; Markov decision process; Bellman equation; Model predictive control; Function (biology)","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":[0.0005827598,0.0002295126,0.0006558171,0.0002116909,0.00008251113,0.00003773974,0.0001045228,0.0001073415,0.00002099675],"category_scores_gemma":[0.0005700911,0.000215057,0.00008134723,0.0003333371,0.00002426371,0.0002748514,0.00001958182,0.0003612697,0.00008414739],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001375373,"about_ca_system_score_gemma":0.00001380448,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001029442,"about_ca_topic_score_gemma":0.00001819122,"domain_scores_codex":[0.9984944,0.000254479,0.0006390905,0.0002707914,0.00009610863,0.0002451326],"domain_scores_gemma":[0.9991615,0.0004201224,0.00009409805,0.0002178703,0.00005312874,0.00005323717],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00004397281,0.000028073,0.003674637,0.0006157625,0.0001558145,0.000005466381,0.0003034351,0.6543461,0.0001508357,0.002403391,0.002411641,0.3358609],"study_design_scores_gemma":[0.006156732,0.0002804786,0.006199914,0.006209722,0.0002958272,0.000007384297,0.0002894409,0.2076132,0.0000277675,0.0008398776,0.7712606,0.0008190367],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002329316,0.06931348,0.9166812,0.0004893105,0.0006565644,0.002404399,0.00006093813,0.0003675356,0.007697243],"genre_scores_gemma":[0.9918704,0.005346475,0.0005900488,0.0004585786,0.0001649222,0.0005753241,0.00004042176,0.00003702667,0.0009167931],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9895411,"threshold_uncertainty_score":0.8769771,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.008009256037607566,"score_gpt":0.2736690119654247,"score_spread":0.2656597559278172,"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."}}