{"id":"W3120459072","doi":"10.18653/v1/2020.wmt-1.110","title":"Improving Parallel Data Identification using Iteratively Refined Sentence Alignments and Bilingual Mappings of Pre-trained Language Models","year":2020,"lang":"en","type":"article","venue":"","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Computer science; Machine translation; Natural language processing; Sentence; Artificial intelligence; Context (archaeology); Metric (unit); Task (project management); Language model; Similarity (geometry)","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003211866,0.0001394447,0.000176022,0.00006837594,0.00007005137,0.0001976838,0.001147144,0.00005351425,0.000002309242],"category_scores_gemma":[0.0001979067,0.0001196379,0.00001945425,0.0003023858,0.0000535926,0.001794821,0.001033836,0.0001003236,5.260169e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001982685,"about_ca_system_score_gemma":0.00007468363,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002146437,"about_ca_topic_score_gemma":0.000006252705,"domain_scores_codex":[0.9985518,0.00004948127,0.0003516871,0.0006031859,0.0002722972,0.0001714783],"domain_scores_gemma":[0.9989363,0.0000401643,0.0002557552,0.0005901741,0.0001027721,0.00007486014],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002082559,0.00002743843,0.00004108845,0.0001559743,0.00001676756,0.00001221945,0.008043975,0.00008528585,0.9618624,0.007899201,0.00005225281,0.02178258],"study_design_scores_gemma":[0.0001740853,0.00003159113,0.00003015029,0.00004814047,0.00000881415,0.000009139136,0.0002188883,0.9144003,0.0827048,0.002227467,0.000004146949,0.0001425036],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07189773,0.0005728253,0.9264146,0.0005577809,0.00002376121,0.0002013358,0.00001762865,0.0002874053,0.00002691338],"genre_scores_gemma":[0.5123009,0.000003788656,0.487393,0.0002350579,0.00001518638,0.000001605052,0.00001349931,0.000005446961,0.0000315423],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.914315,"threshold_uncertainty_score":0.4878692,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06009170999017288,"score_gpt":0.314486357809468,"score_spread":0.2543946478192952,"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."}}