{"id":"W3128846734","doi":"10.4000/ijcol.298","title":"An Exploration of Semantic Features in an Unsupervised Thematic Fit Evaluation Framework","year":2015,"lang":"en","type":"article","venue":"Italian Journal of Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"ca_institutions":"Innovation Cluster (Canada)","funders":"","keywords":"Thematic map; Computer science; Natural language processing; Information retrieval; Artificial intelligence; Cartography; Geography","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":[],"consensus_categories":[],"category_scores_codex":[0.001618012,0.000123635,0.0002436103,0.0003431251,0.00003493839,0.0001531455,0.0006896913,0.00008755548,0.000003331515],"category_scores_gemma":[0.003768595,0.0001092331,0.00004665201,0.0003569648,0.00004298069,0.0009633619,0.00003475248,0.0002527488,0.000001256384],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008916989,"about_ca_system_score_gemma":0.0004426704,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004664153,"about_ca_topic_score_gemma":0.000005142384,"domain_scores_codex":[0.9978127,0.000265179,0.0006480879,0.000150882,0.001000893,0.0001222385],"domain_scores_gemma":[0.9953094,0.0002297333,0.0005766971,0.0002343876,0.003516936,0.0001329028],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001234455,0.0008884107,0.001869229,0.000183291,0.00005200859,0.000128942,0.02608995,0.5305985,0.0002374409,0.408443,0.0003283827,0.03105742],"study_design_scores_gemma":[0.0002937614,0.0003527009,0.0004837472,0.0002700483,0.00001705855,0.00002569152,0.0002502743,0.2822666,0.00046477,0.7154695,0.000006409874,0.00009941685],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05708023,0.000754141,0.9411897,0.0002062353,0.0005077756,0.0001485238,0.000002850177,0.00004100972,0.00006951691],"genre_scores_gemma":[0.5464569,8.786317e-7,0.4533111,0.00004510982,0.0001705919,0.000001180339,0.000007053767,0.000005995871,0.000001204097],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.4893767,"threshold_uncertainty_score":0.4511633,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0878316024480657,"score_gpt":0.3852867768645928,"score_spread":0.2974551744165271,"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."}}