{"id":"W4415731737","doi":"10.1038/s43856-026-01558-x","title":"Interpretable Multiple Instance Learning for Hematologic Diagnosis from Peripheral Blood Smears","year":2025,"lang":"en","type":"preprint","venue":"Communications Medicine","topic":"Digital Imaging for Blood Diseases","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"National Cancer Institute; National Institutes of Health; Memorial Sloan-Kettering Cancer Center","keywords":"Interpretability; Encoder; Pattern recognition (psychology); Feature (linguistics); Function (biology); Pipeline (software); Softmax function; Peripheral blood","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":[0.0008732125,0.001023936,0.0007249788,0.0008546269,0.0002790105,0.0009152047,0.001670866,0.001246702,0.002326415],"category_scores_gemma":[0.002900898,0.0003847653,0.0009012213,0.0005803672,0.0003865949,0.001273112,0.001180084,0.002208763,0.0008259448],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00100251,"about_ca_system_score_gemma":0.0009068431,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005107527,"about_ca_topic_score_gemma":0.006898247,"domain_scores_codex":[0.9996351,0.00008238266,0.00002146917,0.0001360249,0.00006939446,0.00005561737],"domain_scores_gemma":[0.9992834,0.0003564697,0.00007978384,0.0001067832,0.000113947,0.0000595452],"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":[0.0007233956,0.0004241084,0.01676565,0.000266795,0.0002350969,0.0006592954,0.0002137415,0.2922338,0.01726699,0.00585556,0.01791673,0.6474389],"study_design_scores_gemma":[0.00001532587,0.00004647983,0.0009458362,0.00001403682,0.00002398887,0.00009141191,0.00002149295,0.988391,0.004833949,0.004725631,0.0008826242,0.00000826102],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2724258,0.004195876,0.6977783,0.002071325,0.000244777,0.0002017406,0.00290277,0.0153234,0.004855906],"genre_scores_gemma":[0.8676004,0.0006440559,0.119992,0.0006028356,0.0001263975,0.0001202029,0.006527463,0.0001853592,0.004201351],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005107527,"threshold_uncertainty_score":0.01015556,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04791972255597907,"score_gpt":0.3228623583760989,"score_spread":0.2749426358201198,"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."}}