{"id":"W4393786486","doi":"10.5281/zenodo.8274019","title":"SEAO datasets - raw and cleaned datasets","year":2024,"lang":"en","type":"dataset","venue":"PolyPublie (École Polytechnique de Montréal)","topic":"Environmental Monitoring and Data Management","field":"Earth and Planetary Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Raw data; Computer science; Data mining","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":[0.001124065,0.002324068,0.001182035,0.003874938,0.001984271,0.002046653,0.003687277,0.001774033,0.01704029],"category_scores_gemma":[0.004386216,0.0005357178,0.001440156,0.007397252,0.0008879449,0.0008098898,0.001606321,0.001987106,0.01935489],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.008377695,"about_ca_system_score_gemma":0.01343011,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.8266071,"about_ca_topic_score_gemma":0.9200194,"domain_scores_codex":[0.9987888,0.0001268985,0.00008571197,0.000239212,0.000463046,0.0002963184],"domain_scores_gemma":[0.9971855,0.0002539234,0.0001277802,0.0005004028,0.001587413,0.0003449579],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00006767258,0.00005614863,0.002512117,0.0002844651,0.00006314887,0.00005404372,0.00003881306,0.001351756,0.0001814015,0.0006372144,0.9911605,0.003592964],"study_design_scores_gemma":[0.0002598923,0.00002731354,0.02341292,0.0003144836,0.00006059231,0.00011132,0.0003319969,0.004011543,0.0008048667,0.001566689,0.9690062,0.00009221082],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.0009946196,0.0001310851,0.0001507871,0.0001014798,0.00003464428,0.00003205243,0.9969047,0.0005437594,0.001106861],"genre_scores_gemma":[0.00122692,0.00005546093,0.0004413021,0.00002859819,0.000004833848,0.00004268206,0.9974309,0.00006479632,0.0007045846],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.8266071,"threshold_uncertainty_score":0.348828,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.009881952169962108,"score_gpt":0.2214231979050645,"score_spread":0.2115412457351024,"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."}}