{"id":"W4388685754","doi":"10.48550/arxiv.2311.07565","title":"Exploration via linearly perturbed loss minimisation","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Advanced Bandit Algorithms Research","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Regret; Minimisation (clinical trials); Perturbation (astronomy); Mathematics; Mathematical optimization; Applied mathematics; Computer science; Linear programming; Simple (philosophy); Algorithm; Statistics; Physics","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00172512,0.0003599336,0.0004871711,0.001056941,0.0002913122,0.0003722809,0.001938301,0.0004699841,0.000325128],"category_scores_gemma":[0.002275386,0.0003562872,0.0003103127,0.00201855,0.0002747666,0.001128648,0.001684637,0.0008737777,0.003949187],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003855352,"about_ca_system_score_gemma":0.0002921142,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001759416,"about_ca_topic_score_gemma":0.000176334,"domain_scores_codex":[0.9958016,0.0004591355,0.0005589002,0.001801097,0.0008825313,0.0004967886],"domain_scores_gemma":[0.9949974,0.001426082,0.0004955607,0.001668555,0.001146759,0.0002656485],"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":[0.0002225817,0.0001157487,0.002442799,0.0000333811,0.00008657186,0.0007160993,0.001021574,0.9804346,0.00009016768,0.00136791,0.004686037,0.008782529],"study_design_scores_gemma":[0.0005115322,0.00007030285,0.002540959,0.00003924666,0.00003133246,0.000002733988,0.0007905017,0.6875276,0.0002581885,0.3045157,0.003270266,0.0004415768],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1145592,0.00002444147,0.8820753,0.0008098495,0.001246683,0.0005399146,0.00006832813,0.0003013771,0.0003749319],"genre_scores_gemma":[0.9643141,0.0001936863,0.0008375053,0.00004986145,0.0003804411,0.000004715367,0.0001238232,0.00005108388,0.03404479],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8812378,"threshold_uncertainty_score":0.9998889,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4737798012038305,"score_gpt":0.3347368947070564,"score_spread":0.1390429064967741,"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."}}