{"id":"W2250491265","doi":"10.63317/24i436sxdny8","title":"Texto4Science: a Quebec French Database of Annotated Short Text Messages","year":2012,"lang":"en","type":"article","venue":"","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":46,"is_retracted":false,"has_abstract":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Period (music); Computer science; Phenomenon; Database; World Wide Web; Library science; Art","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009183222,0.002292968,0.001097192,0.01140284,0.002150917,0.002004364,0.002240851,0.001859555,0.02506322],"category_scores_gemma":[0.007874903,0.0005140699,0.0008121402,0.01003249,0.0006685742,0.001971741,0.0009749375,0.001040641,0.01391715],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005495195,"about_ca_system_score_gemma":0.01158089,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.6333551,"about_ca_topic_score_gemma":0.6296532,"domain_scores_codex":[0.9987447,0.0002212868,0.0001070759,0.0003134567,0.0004721964,0.0001412763],"domain_scores_gemma":[0.9953107,0.001097274,0.0002511237,0.0004989973,0.002501062,0.000340784],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.001238692,0.0002217921,0.006369815,0.002884062,0.0001912109,0.0009605575,0.0008779741,0.003431962,0.01041595,0.004145843,0.8719248,0.09733732],"study_design_scores_gemma":[0.0004291314,0.0001248768,0.03491121,0.0005113333,0.0002308656,0.0005250852,0.001164856,0.02209317,0.01092878,0.002227488,0.9266265,0.0002268018],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.01991213,0.001718462,0.005588885,0.0006478563,0.0001331446,0.0004082576,0.9505461,0.01231313,0.008731963],"genre_scores_gemma":[0.0272724,0.00073381,0.00991478,0.0001492387,0.00004424195,0.0003833598,0.9530153,0.0008187389,0.007668124],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.3666449,"threshold_uncertainty_score":0.7376083,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01912972885074295,"score_gpt":0.3000327276038703,"score_spread":0.2809029987531274,"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."}}