{"id":"W2903182367","doi":"10.18653/v1/w18-6480","title":"Measuring sentence parallelism using Mahalanobis distances: The NRC unsupervised submissions to the WMT18 Parallel Corpus Filtering shared task","year":2018,"lang":"en","type":"article","venue":"","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Computer science; Natural language processing; Task (project management); Artificial intelligence; Sample (material); Word (group theory); Sentence; Parallelism (grammar); Speech recognition; Linguistics; Parallel computing","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":[0.01620418,0.003606248,0.002234631,0.003929873,0.003155218,0.003962121,0.003270987,0.003444179,0.008341022],"category_scores_gemma":[0.08477108,0.001047494,0.002014055,0.003550501,0.001425688,0.004247446,0.00795058,0.00515909,0.009981167],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00175115,"about_ca_system_score_gemma":0.003859224,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006871508,"about_ca_topic_score_gemma":0.01488423,"domain_scores_codex":[0.9779375,0.009440991,0.002017915,0.00499182,0.004666046,0.0009456946],"domain_scores_gemma":[0.9380894,0.02693183,0.001706513,0.0141777,0.01639429,0.002700263],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.003478767,0.001520134,0.01717103,0.002935666,0.001228444,0.0009253456,0.004328132,0.01865239,0.05850955,0.003628696,0.3858809,0.501741],"study_design_scores_gemma":[0.002542552,0.00387658,0.08684748,0.0006305628,0.0009034948,0.002285884,0.005440111,0.4126451,0.150411,0.03780163,0.2953673,0.001248383],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.4684553,0.003235151,0.3413064,0.003369701,0.004417259,0.003328468,0.08368497,0.06294405,0.02925864],"genre_scores_gemma":[0.4752846,0.0003732925,0.3301123,0.001120858,0.0006386474,0.004535564,0.1639575,0.01057774,0.0133994],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.01620418,"threshold_uncertainty_score":0.08569694,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05648903696019823,"score_gpt":0.2793418193547273,"score_spread":0.2228527823945291,"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."}}