{"id":"W3009074285","doi":"10.1002/ece3.6147","title":"Three critical factors affecting automated image species recognition performance for camera traps","year":2020,"lang":"en","type":"article","venue":"Ecology and Evolution","topic":"Species Distribution and Climate Change","field":"Environmental Science","cited_by":157,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute; University of Calgary; University of Guelph","funders":"","keywords":"Artificial intelligence; Computer science; Deep learning; Transfer of learning; Camera trap; Pattern recognition (psychology); Machine learning; Computer vision; Wildlife; Ecology","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":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00009454825,0.00008397554,0.00009805061,0.00001265258,0.000303543,0.00001758895,0.0000431135,0.00008636204,0.006693633],"category_scores_gemma":[0.0002180907,0.00008036464,0.00003170038,0.00009416558,0.0002188298,0.0002137722,0.00003726512,0.00007666655,0.0003013797],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001845253,"about_ca_system_score_gemma":0.000004319555,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002292445,"about_ca_topic_score_gemma":0.0002605996,"domain_scores_codex":[0.9994018,0.00002050794,0.0001071982,0.0001961724,0.00005906579,0.0002152396],"domain_scores_gemma":[0.9997383,0.0001008079,0.00003236926,0.00003915664,0.00001254077,0.00007681797],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001455926,0.0001221734,0.9170343,0.0001136406,0.00001811149,0.000002738919,0.001197442,0.00001251442,0.04844391,0.0007314058,0.03068176,0.001496354],"study_design_scores_gemma":[0.0002651718,0.0002293776,0.9886783,0.000003124801,0.00001552899,0.000004236342,0.00108869,0.007620837,0.001113072,0.0001568377,0.000721916,0.0001028729],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9947466,0.00001485419,0.001284238,0.001247595,0.0001659315,0.0001783779,0.00006906128,0.0001703164,0.002123065],"genre_scores_gemma":[0.9993991,0.00001628079,0.0002077429,0.0001911261,0.00004800877,0.00001694698,0.0001016352,0.000005346106,0.00001377569],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.07164398,"threshold_uncertainty_score":0.9942144,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04019360546685947,"score_gpt":0.2579962134461807,"score_spread":0.2178026079793213,"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."}}