{"id":"W4220808180","doi":"10.1016/j.isprsjprs.2022.02.023","title":"Spatiotemporal distribution of labeled data can bias the validation and selection of supervised learning algorithms: A marine remote sensing example","year":2022,"lang":"en","type":"article","venue":"ISPRS Journal of Photogrammetry and Remote Sensing","topic":"Marine and coastal ecosystems","field":"Earth and Planetary Sciences","cited_by":31,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Algorithm; Computer science; Data set; Remote sensing; Satellite; Matching (statistics); Sampling (signal processing); Set (abstract data type); Data mining; Statistics; Mathematics; Artificial intelligence; 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":[0.001974213,0.0001216939,0.0003033581,0.0001429667,0.0003739769,0.0000573586,0.000114051,0.00004411208,0.00003907631],"category_scores_gemma":[0.0001920972,0.00009283952,0.00005281026,0.0005313351,0.00007300128,0.0001503674,0.0001125466,0.0003844595,8.885168e-8],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001123719,"about_ca_system_score_gemma":0.00007655034,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.1203861,"about_ca_topic_score_gemma":0.0093728,"domain_scores_codex":[0.9982049,0.0005098099,0.0005535408,0.0001811404,0.0003739705,0.0001765874],"domain_scores_gemma":[0.9986841,0.0002528908,0.000682255,0.00013665,0.0001693915,0.00007473054],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001380927,0.00000643715,0.01012427,0.00006396796,0.00005453414,0.00001347675,0.0002846223,0.0007839657,0.00236602,5.444903e-7,0.0000157038,0.9861484],"study_design_scores_gemma":[0.0006335381,0.0004965914,0.01326739,0.0000870042,0.0001046128,0.001270837,0.001941606,0.9753154,0.004695421,0.0001856129,0.001866891,0.0001351062],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9789857,0.0001846105,0.02026922,0.0001231992,0.0001749863,0.0001310677,0.00005312184,0.000007822838,0.00007030474],"genre_scores_gemma":[0.9939767,0.0001541274,0.00544893,0.00002016088,0.00009882432,2.038277e-9,0.0002754106,0.000005259141,0.00002065088],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9860132,"threshold_uncertainty_score":0.8854713,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03792260276340074,"score_gpt":0.2373618406598871,"score_spread":0.1994392378964863,"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."}}