{"id":"W2901460997","doi":"10.1016/j.envint.2018.11.042","title":"A picture tells a thousand…exposures: Opportunities and challenges of deep learning image analyses in exposure science and environmental epidemiology","year":2018,"lang":"en","type":"article","venue":"Environment International","topic":"Health, Environment, Cognitive Aging","field":"Environmental Science","cited_by":105,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto; University of British Columbia; McGill University; McGill University Health Centre","funders":"Fonds de Recherche du Québec - Santé; British Columbia Lung Association; Natural Sciences and Engineering Research Council of Canada; Wellcome Trust; Canadian Institutes of Health Research; Cancer Research Society","keywords":"Environmental epidemiology; Epidemiology; Exposure assessment; Environmental health; Data science; Computer science; Medicine; Pathology","routes":{"ca_aff":true,"ca_fund":true,"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":["sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.001782364,0.0002531408,0.0003133099,0.0002100364,0.0002126895,0.00002079375,0.0003213816,0.0001071472,0.002013508],"category_scores_gemma":[0.0003455342,0.0002421261,0.00003833339,0.00007565076,0.00360853,0.0004934216,0.0006718773,0.0002686484,0.00005134616],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003426598,"about_ca_system_score_gemma":0.00001152601,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007058629,"about_ca_topic_score_gemma":0.00003079614,"domain_scores_codex":[0.9973053,0.0002962087,0.0005106085,0.0008235282,0.0006236144,0.0004407379],"domain_scores_gemma":[0.9989482,0.000326077,0.0002681679,0.0002583358,0.000005538675,0.0001937407],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001004197,0.0003750178,0.629764,0.00003914005,0.00007884728,0.00004510537,0.008351056,0.0007365198,0.1531295,0.0005192309,0.00002935232,0.2068318],"study_design_scores_gemma":[0.0007602184,0.0004561313,0.96775,0.00004290088,0.00002514586,0.00005629267,0.005851699,0.004051911,0.005415898,0.0005106009,0.0147285,0.000350719],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9902366,0.001643842,0.0007437915,0.001299614,0.00007309645,0.0002682675,0.00002384843,0.0000172102,0.005693667],"genre_scores_gemma":[0.9784505,0.01851887,0.002430003,0.0003688273,0.00006857216,0.00002448952,0.00001691464,0.00002205993,0.00009972516],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3379861,"threshold_uncertainty_score":0.9991031,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0776304045609843,"score_gpt":0.3232593032122717,"score_spread":0.2456288986512874,"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."}}