{"id":"W7102329764","doi":"","title":"Offline Preference Optimization via Maximum Marginal Likelihood Estimation","year":2025,"lang":"","type":"article","venue":"ArXiv.org","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Hyperparameter; Preference; Marginal likelihood; Entropy (arrow of time); Principle of maximum entropy; Revealed preference","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"],"consensus_categories":[],"category_scores_codex":[0.0007621924,0.0004903995,0.0005236161,0.0004076937,0.0004235751,0.0004443146,0.001331448,0.0003369727,0.0001338222],"category_scores_gemma":[0.00008991516,0.0005014415,0.0001642703,0.00132434,0.00009022216,0.001333398,0.0006789247,0.0004450517,0.0001613836],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002182815,"about_ca_system_score_gemma":0.0003368374,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002392301,"about_ca_topic_score_gemma":0.00003018541,"domain_scores_codex":[0.9965068,0.0002746975,0.001042614,0.001075578,0.0004181913,0.0006821854],"domain_scores_gemma":[0.9975038,0.0001060416,0.0004701256,0.001322592,0.000421204,0.0001762358],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00009644698,0.001091699,0.1968583,0.0009173435,0.000307987,0.00004151868,0.001004918,0.01870153,0.0009838511,0.01114171,0.01141168,0.757443],"study_design_scores_gemma":[0.0006437735,0.0002759501,0.0410698,0.0006732861,0.00006674731,0.00001976194,0.00001884024,0.9422321,0.004670158,0.005867617,0.003940515,0.0005214814],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0144416,0.0007169536,0.9709091,0.004915839,0.002083479,0.0008438583,0.000006472772,0.0004769524,0.00560578],"genre_scores_gemma":[0.8689339,0.0002130371,0.1285922,0.0005858723,0.0001475795,0.0001002001,0.00002961736,0.00002499672,0.001372603],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9235305,"threshold_uncertainty_score":0.9997437,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03695703122362177,"score_gpt":0.2691590736295219,"score_spread":0.2322020424059001,"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."}}