{"id":"W4380871602","doi":"10.1093/bioinformatics/btad390","title":"Interpretable deep learning architectures for improving drug response prediction performance: myth or reality?","year":2023,"lang":"en","type":"article","venue":"Bioinformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":30,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute; McGill University","funders":"Natural Sciences and Engineering Research Council of Canada; Institut de Valorisation des Données; Merck Canada; Canada First Research Excellence Fund; Government of Canada; Génome Québec; Compute Canada","keywords":"Interpretability; Computer science; Generalizability theory; Machine learning; Artificial intelligence; Baseline (sea); Random forest; Deep learning; Overfitting; Perceptron; Predictive modelling; Data mining; Artificial neural network; Statistics; Mathematics","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":[],"consensus_categories":[],"category_scores_codex":[0.002594439,0.0001702498,0.0001860833,0.000360592,0.0003470672,0.0002671624,0.0005833443,0.00005704517,0.000005054613],"category_scores_gemma":[0.001118576,0.0001398539,0.00008686463,0.0007612993,0.00004040548,0.0006087423,0.0004006201,0.0002079645,0.00004023294],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001033496,"about_ca_system_score_gemma":0.0002149045,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000006343452,"about_ca_topic_score_gemma":0.000003144507,"domain_scores_codex":[0.9983529,0.0001621967,0.0004990714,0.0002139039,0.0003751983,0.0003967159],"domain_scores_gemma":[0.9978152,0.001407423,0.0002055518,0.0003668994,0.0001143268,0.00009058126],"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":[0.0004884016,0.0000161731,0.0001738307,0.0003855272,0.00002113258,0.000001963395,0.01198546,0.6252865,0.0001182072,0.0004948928,0.001308893,0.3597191],"study_design_scores_gemma":[0.0003448644,0.0002085064,0.003048693,0.00004873926,0.000008462102,0.00002053769,0.0004185759,0.9902318,0.0005533557,0.0007253331,0.004222837,0.0001683321],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2583678,0.00001253791,0.7394581,0.0003742476,0.0004602529,0.0003519313,0.00001568125,0.0006000636,0.0003593916],"genre_scores_gemma":[0.612896,0.00001390259,0.3847844,0.0003227455,0.0001408117,0.0001177965,0.0000536952,0.0000258246,0.001644909],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.3649453,"threshold_uncertainty_score":0.5703077,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02475005689958606,"score_gpt":0.2980367158105436,"score_spread":0.2732866589109576,"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."}}