{"id":"W4386767067","doi":"10.1101/2023.09.11.557208","title":"Tissue-specific transfer-learning enables retasking of a general comprehensive model to a specific domain","year":2023,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Molecular Biology Techniques and Applications","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Alberta Children's Hospital; Hotchkiss Brain Institute; University of Calgary","funders":"Natural Sciences and Engineering Research Council of Canada; National Institutes of Health; Alberta Innovates","keywords":"Computer science; Flexibility (engineering); Transfer of learning; Artificial intelligence; Machine learning; Protocol (science); Annotation; Sample (material); Domain (mathematical analysis); Training set; Medicine; Statistics; Mathematics; 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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002711874,0.001376474,0.0008686135,0.0004485244,0.0004522906,0.0009699881,0.002081537,0.001425286,0.003017548],"category_scores_gemma":[0.007033822,0.0004794398,0.001338793,0.0004217775,0.001504148,0.001805405,0.003768411,0.003647673,0.001436529],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008479591,"about_ca_system_score_gemma":0.001360323,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002383571,"about_ca_topic_score_gemma":0.002627456,"domain_scores_codex":[0.9992059,0.0002628733,0.00003854596,0.0002803328,0.000115443,0.00009686665],"domain_scores_gemma":[0.9974849,0.0009996769,0.0001272265,0.0009152496,0.0002641452,0.0002088326],"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":[0.0005539954,0.000396782,0.006789493,0.0002388062,0.0002834288,0.0004533001,0.0002466303,0.7393273,0.07315639,0.01095353,0.009112345,0.158488],"study_design_scores_gemma":[0.00001845903,0.0001120719,0.0004170722,0.00001033458,0.00001808614,0.00004350401,0.00001829909,0.9730895,0.01666897,0.007881626,0.001703688,0.00001860673],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1715077,0.0003234255,0.8162113,0.0008838061,0.0001711213,0.0001865006,0.0007363499,0.007352121,0.002627583],"genre_scores_gemma":[0.8105621,0.0001818466,0.1804256,0.0006740841,0.00009282014,0.0003599199,0.002273974,0.0009470959,0.004482445],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003017548,"threshold_uncertainty_score":0.01434195,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02569501127095022,"score_gpt":0.253542861008352,"score_spread":0.2278478497374018,"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."}}