{"id":"W4389519548","doi":"10.18653/v1/2023.wmt-1.51","title":"Results of WMT23 Metrics Shared Task: Metrics Might Be Guilty but References Are Not Innocent","year":2023,"lang":"en","type":"article","venue":"","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"Bundesministerium für Bildung und Forschung; Deutsche Forschungsgemeinschaft","keywords":"George (robot); Task (project management); Computer science; Machine translation; Artificial intelligence; Engineering; Systems engineering","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02037994,0.005040739,0.002906356,0.009366162,0.004520294,0.006057801,0.003362204,0.005655712,0.0204068],"category_scores_gemma":[0.09116952,0.0007894732,0.002741561,0.005209699,0.001777417,0.007811032,0.008688097,0.003336666,0.02489996],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002980336,"about_ca_system_score_gemma":0.004385125,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.03419778,"about_ca_topic_score_gemma":0.06814105,"domain_scores_codex":[0.9676339,0.01417267,0.002911726,0.00428996,0.00926548,0.001726372],"domain_scores_gemma":[0.9319125,0.02469779,0.002756431,0.01705656,0.01822443,0.005352273],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"observational","study_design_scores_codex":[0.001217552,0.0004885866,0.005627603,0.001138953,0.0004757126,0.0004031886,0.000523224,0.00242026,0.001790312,0.001879459,0.9132499,0.07078524],"study_design_scores_gemma":[0.002810856,0.001968744,0.04599922,0.0012304,0.0009137338,0.00204161,0.003951402,0.06118887,0.01998228,0.0361357,0.8229981,0.0007790632],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"empirical","genre_scores_codex":[0.2223986,0.0190029,0.03748726,0.02363833,0.01896982,0.002100246,0.4778467,0.0915569,0.1069993],"genre_scores_gemma":[0.2820669,0.0009505452,0.0404283,0.004472195,0.001467467,0.001416242,0.6195403,0.01354893,0.03610902],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.03419778,"threshold_uncertainty_score":0.1077806,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07787230308748376,"score_gpt":0.3156563963070204,"score_spread":0.2377840932195366,"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."}}