{"id":"W2111141593","doi":"10.1109/iccv.2007.4408877","title":"Learning Structured Appearance Models from Captioned Images of Cluttered Scenes","year":2007,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Artificial intelligence; Feature (linguistics); ENCODE; Computer vision; Object (grammar); Graph; Pattern recognition (psychology)","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009485485,0.001890596,0.001332636,0.001581953,0.0004539531,0.00147184,0.002705523,0.002422211,0.001809559],"category_scores_gemma":[0.005081295,0.001072975,0.001542372,0.001574667,0.0009870944,0.00214273,0.0008935236,0.002547318,0.001299709],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009585731,"about_ca_system_score_gemma":0.0004727222,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004229298,"about_ca_topic_score_gemma":0.00762704,"domain_scores_codex":[0.999202,0.00021467,0.00002651103,0.0003786798,0.0001055347,0.00007263076],"domain_scores_gemma":[0.9972677,0.001321125,0.000263947,0.0005672835,0.0004239024,0.0001560818],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001007646,0.0006215411,0.006069622,0.0004828791,0.0003627713,0.001176519,0.0005765747,0.5668203,0.04114448,0.004919944,0.01918442,0.3576334],"study_design_scores_gemma":[0.00001547416,0.00004927183,0.00047536,0.00001012459,0.00001670122,0.00009379033,0.0000350542,0.9946809,0.002083264,0.001998704,0.0005305634,0.00001061776],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1285521,0.0004951116,0.8616598,0.000314639,0.0001266676,0.0002486134,0.001176164,0.005499383,0.00192757],"genre_scores_gemma":[0.6277601,0.0004592928,0.3592835,0.0004026198,0.0001655483,0.0003243777,0.00772335,0.0005300039,0.00335123],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004229298,"threshold_uncertainty_score":0.008409381,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01213481241210248,"score_gpt":0.2609122937360986,"score_spread":0.2487774813239962,"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."}}