{"id":"W7126165022","doi":"10.21428/594757db.b52553ec","title":"Taxonomic Reasoning for Rare Arthropods: Combining Dense Image Captioning and RAG for Interpretable Classification","year":2025,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Guelph","funders":"","keywords":"Context (archaeology); Biodiversity; Closed captioning; Taxonomic rank; Matching (statistics); Interpretability; Contextual image classification; Convolutional neural network","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":[0.001318746,0.001051214,0.0004785239,0.002173698,0.0004142808,0.001618391,0.001636567,0.001450203,0.003649418],"category_scores_gemma":[0.005517894,0.0002952959,0.001164065,0.0008171265,0.0008019644,0.002863896,0.001563786,0.001492307,0.001885545],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008175319,"about_ca_system_score_gemma":0.0006762181,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004046194,"about_ca_topic_score_gemma":0.005192006,"domain_scores_codex":[0.9993724,0.0001647821,0.00003500627,0.0002191634,0.0001359209,0.00007267114],"domain_scores_gemma":[0.998229,0.0007654638,0.0001922929,0.0003919515,0.0003333353,0.00008807541],"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":[0.0004408068,0.0002520407,0.004046271,0.0004782933,0.00009542867,0.0007172227,0.0006249321,0.05357539,0.06285578,0.009415222,0.02156343,0.8459353],"study_design_scores_gemma":[0.00002890413,0.0001266536,0.00225599,0.0001106756,0.00006590565,0.0003893757,0.0002761114,0.9250223,0.02781754,0.02975192,0.01410272,0.00005202048],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06867969,0.001946823,0.8964963,0.002099189,0.0003711765,0.0003506301,0.002182524,0.02054086,0.007332864],"genre_scores_gemma":[0.4316138,0.0006686199,0.5580639,0.001207232,0.0002453774,0.0001760561,0.004524822,0.0006031056,0.002897017],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004046194,"threshold_uncertainty_score":0.01220852,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01547139791379805,"score_gpt":0.3021148857597908,"score_spread":0.2866434878459928,"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."}}