{"id":"W2807690984","doi":"10.1515/cllt-2017-0031","title":"Dependency profiles in the large-scale analysis of discourse connectives","year":2018,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University; Brock University","funders":"Turun Yliopisto","keywords":"Dependency (UML); Computer science; Syntax; Natural language processing; Scope (computer science); Artificial intelligence; Linguistics; Cluster analysis; Focus (optics); Annotation","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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.001492473,0.0001660664,0.000317524,0.0003305053,0.0001608396,0.00009942574,0.00089115,0.00007872504,0.00001091374],"category_scores_gemma":[0.009501935,0.0001139119,0.00007580323,0.0009034792,0.0003694422,0.00002124211,0.0002372914,0.0002139225,0.000001529016],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001666939,"about_ca_system_score_gemma":0.0000791128,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004682712,"about_ca_topic_score_gemma":0.0001326414,"domain_scores_codex":[0.9985777,0.0001964655,0.0003387568,0.000377119,0.0002380456,0.0002719246],"domain_scores_gemma":[0.9977282,0.000931224,0.0002060834,0.0005922937,0.0004916548,0.00005053272],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00001931686,0.000085678,0.003078181,0.00002643428,0.00005954867,0.000026806,0.009592428,0.00000146185,0.00006619215,0.9840117,0.00001252895,0.003019712],"study_design_scores_gemma":[0.0002634989,0.0002042318,0.003194904,0.0001128853,0.000390102,0.0000076438,0.000977191,0.02266317,0.002030306,0.9684435,0.00139966,0.0003128761],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05592057,0.007471159,0.9158127,0.00009160334,0.001373527,0.0004604826,0.00008027069,0.0002714368,0.01851822],"genre_scores_gemma":[0.9420388,0.00002496709,0.05728822,0.0001337012,0.0004336173,0.00001043824,0.000006283097,0.000009197206,0.00005472481],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8861183,"threshold_uncertainty_score":0.9988415,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.00909766402107549,"score_gpt":0.2919299757573142,"score_spread":0.2828323117362387,"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."}}