{"id":"W4388928478","doi":"10.1093/bioadv/vbad166","title":"Adversarial training improves model interpretability in single-cell RNA-seq analysis","year":2023,"lang":"en","type":"article","venue":"Bioinformatics Advances","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"Princess Margaret Cancer Centre; Lunenfeld-Tanenbaum Research Institute; University Health Network; University of Toronto; Sinai Health System; University of Waterloo","funders":"","keywords":"Interpretability; Robustness (evolution); Machine learning; Computer science; Artificial intelligence; Classifier (UML); Data mining; Biology; Gene","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003012273,0.0002049645,0.0002933449,0.0002249773,0.00006713717,0.00004771689,0.0002664299,0.0001462825,0.000005960308],"category_scores_gemma":[0.00009753006,0.000193964,0.0002130896,0.0005989606,0.00009526016,0.00003839514,0.00008761912,0.0001171016,0.0000107046],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002938096,"about_ca_system_score_gemma":0.0000785763,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001880012,"about_ca_topic_score_gemma":0.0002667725,"domain_scores_codex":[0.9986461,0.0000253361,0.0005237252,0.0002611091,0.000173332,0.0003703895],"domain_scores_gemma":[0.9993692,0.00003043663,0.0001406749,0.0003253566,0.00005066741,0.00008367641],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0004164279,0.0003055869,0.006722047,0.0003336331,0.0002998674,0.000004438248,0.006948936,0.1758245,0.735035,0.00009353106,0.0002098065,0.07380629],"study_design_scores_gemma":[0.00120014,0.0002856604,0.0004635194,0.00002338191,0.0001314543,0.000001112419,0.002163432,0.920497,0.07108972,0.0004048858,0.00322427,0.0005153905],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.949392,0.0002771257,0.04497312,0.00004839431,0.0003278592,0.0002075745,0.00006554185,0.00006266032,0.00464569],"genre_scores_gemma":[0.9874762,0.000241751,0.01160591,0.0001161604,0.00007320147,0.00001670274,0.000235716,0.00001685676,0.0002175463],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7446725,"threshold_uncertainty_score":0.7909622,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02329046184857348,"score_gpt":0.2499431431405594,"score_spread":0.2266526812919859,"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."}}