{"id":"W3174355057","doi":"10.1609/aaai.v35i11.17163","title":"DIBS: Diversity Inducing Information Bottleneck in Model Ensembles","year":2021,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal; Vector Institute; University of Toronto","funders":"","keywords":"MNIST database; Computer science; Generalization; Machine learning; Artificial intelligence; Benchmark (surveying); Information bottleneck method; Bayesian probability; Bottleneck; Artificial neural network; Data mining; Mutual information; Mathematics; Geography","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":[0.0005427068,0.0001725914,0.0002223386,0.0001610763,0.0002953961,0.0002324181,0.001586316,0.0001016285,0.00002157672],"category_scores_gemma":[0.001038655,0.0001508067,0.00008122849,0.000840952,0.0001079526,0.001615345,0.00191551,0.0004542001,0.00003360652],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009039568,"about_ca_system_score_gemma":0.0001742507,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009138243,"about_ca_topic_score_gemma":0.00002870185,"domain_scores_codex":[0.9983438,0.0000260717,0.0004778917,0.000338508,0.0005185571,0.0002951522],"domain_scores_gemma":[0.9986626,0.00008698904,0.0003317318,0.0002772297,0.0005836858,0.00005775532],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002739183,0.00007461213,0.002157912,0.00004151733,0.000007649981,0.000001233922,0.006551681,0.04395501,0.006608524,0.8967974,0.00003567826,0.04374143],"study_design_scores_gemma":[0.00003455048,0.00002737979,0.000536027,0.0001769556,0.00000511642,0.00000297892,0.0008882771,0.7343059,0.121793,0.1420485,0.00001452943,0.0001668118],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4885107,0.00001488532,0.4928103,0.003680272,0.0004669024,0.0002625519,0.000003203123,0.0001088754,0.01414225],"genre_scores_gemma":[0.9855833,0.00002044263,0.0140419,0.0002501625,0.00002450407,0.000006006664,6.465206e-7,0.00000530249,0.00006772417],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7547489,"threshold_uncertainty_score":0.614972,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07016165741512634,"score_gpt":0.2809434166131086,"score_spread":0.2107817591979823,"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."}}