{"id":"W3087570018","doi":"10.1002/ece3.6692","title":"Improving the accessibility and transferability of machine learning algorithms for identification of animals in camera trap images: MLWIC2","year":2020,"lang":"en","type":"article","venue":"Ecology and Evolution","topic":"Species Distribution and Climate Change","field":"Environmental Science","cited_by":78,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Saskatchewan; Ministry of Natural Resources and Forestry","funders":"","keywords":"Camera trap; Artificial intelligence; Computer science; Transferability; Sample (material); Identification (biology); Machine learning; Trap (plumbing); Deep learning; Animal species; Convolutional neural network; Software; Range (aeronautics); Filter (signal processing); Wildlife; Computer vision; Ecology; Geography; Biology","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":[],"consensus_categories":[],"category_scores_codex":[0.0005049026,0.00004609513,0.0001074017,0.000009502267,0.00008072941,0.00000478076,0.00005123794,0.00004690723,0.0003658998],"category_scores_gemma":[0.0001734207,0.00003713613,0.00002089145,0.00009474668,0.0002608381,0.00009180555,0.00002968294,0.0000677702,0.000001149475],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006908684,"about_ca_system_score_gemma":0.000005015191,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002882809,"about_ca_topic_score_gemma":0.0005182537,"domain_scores_codex":[0.999411,0.00007705879,0.0002149087,0.0001666939,0.00004500731,0.00008528189],"domain_scores_gemma":[0.9997539,0.00008130731,0.00008177693,0.00005012511,0.00001074122,0.00002214567],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00009313122,0.00004768399,0.8305702,0.00007041105,0.000003470964,7.400585e-8,0.0008463068,0.00004071484,0.1649135,0.0001643956,0.00001451835,0.003235658],"study_design_scores_gemma":[0.0003081659,0.0001198664,0.9789514,0.000001201693,0.000009830334,6.024922e-7,0.0009038334,0.01406686,0.005367538,0.0002144958,0.00002283317,0.00003338663],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9959626,0.0001720568,0.002728146,0.0007679513,0.00002036305,0.000239944,0.00005769781,0.000006275437,0.00004493323],"genre_scores_gemma":[0.9998505,0.00004590588,0.00004175527,0.00002105608,0.000004981776,0.00001832487,0.00001114128,0.000001848305,0.000004450997],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1595459,"threshold_uncertainty_score":0.4006347,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02621835201825202,"score_gpt":0.2657396341631272,"score_spread":0.2395212821448752,"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."}}