{"id":"W4282827306","doi":"10.1109/icse-companion55297.2022.9793750","title":"HUDD: A tool to debug DNNs for safety analysis","year":2022,"lang":"en","type":"article","venue":"2022 IEEE/ACM 44th International Conference on Software Engineering: Companion Proceedings (ICSE-Companion)","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Ottawa","funders":"European Research Council; Natural Sciences and Engineering Research Council of Canada; Fonds National de la Recherche Luxembourg; European Commission","keywords":"Debugging; Computer science; Cluster analysis; Retraining; Root cause; Deep neural networks; Root (linguistics); Artificial neural network; Artificial intelligence; Image (mathematics); Domain (mathematical analysis); Pattern recognition (psychology); Machine learning; Reliability engineering; 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":["metaepi_narrow","open_science","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.001346175,0.0007946675,0.0009491141,0.001792629,0.0009041613,0.0007603216,0.005616846,0.0001568177,0.001060446],"category_scores_gemma":[0.002075682,0.0009244552,0.0005200035,0.003316492,0.00004792857,0.0009940165,0.002093493,0.001326607,0.00009195142],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001068034,"about_ca_system_score_gemma":0.0002631834,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004787123,"about_ca_topic_score_gemma":0.000004421393,"domain_scores_codex":[0.9939189,0.00007104828,0.001182318,0.001744392,0.002095943,0.0009874076],"domain_scores_gemma":[0.9962902,0.000559847,0.0005720584,0.000886762,0.001273421,0.0004176956],"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":[0.000195104,0.0002969842,0.008803791,0.00007043211,0.0008107013,0.00001534,0.001818738,0.4829992,0.001285356,0.4954883,0.005476744,0.002739302],"study_design_scores_gemma":[0.001363999,0.0004104444,0.008859153,0.00007946401,0.0001506585,0.00004849808,0.0003882661,0.9384956,0.0003608436,0.001761426,0.04677783,0.001303817],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0437402,0.00002082122,0.9457972,0.002519684,0.004605706,0.001039809,0.0002441839,0.001446569,0.0005858392],"genre_scores_gemma":[0.8496048,0.00001039537,0.1469738,0.0008831674,0.0005354157,0.0007036959,0.0002794154,0.0001050069,0.0009043711],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8058646,"threshold_uncertainty_score":0.9998527,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02880629283998286,"score_gpt":0.278955330687586,"score_spread":0.2501490378476032,"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."}}