{"id":"W2972204217","doi":"10.1109/icsme.2019.00078","title":"DeepEvolution: A Search-Based Testing Approach for Deep Neural Networks","year":2019,"lang":"en","type":"preprint","venue":"","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Fuzz testing; Computer science; Machine learning; Artificial intelligence; Code coverage; Artificial neural network; Random testing; Deep neural networks; Model-based testing; Training set; Test case; Data mining; Software; Programming language","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001111189,0.00046812,0.0005137089,0.0002133667,0.0003164904,0.000498522,0.002786229,0.0004646754,0.00001133452],"category_scores_gemma":[0.0005140155,0.000451477,0.0002764304,0.0004815702,0.00008950435,0.0002657564,0.003157848,0.001649508,0.000008424957],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002117765,"about_ca_system_score_gemma":0.0002975738,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001194034,"about_ca_topic_score_gemma":0.000004216625,"domain_scores_codex":[0.9965439,0.0002809252,0.0004573402,0.001397771,0.0005083664,0.0008117413],"domain_scores_gemma":[0.9966223,0.001010236,0.0002873023,0.00154835,0.0003732037,0.0001586468],"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":[0.00001322078,0.00003744184,0.002071934,0.0001839665,0.00002378915,0.000002965975,0.0000617209,0.9684583,0.000001601757,0.004417406,0.00009169926,0.02463595],"study_design_scores_gemma":[0.0005473308,0.0000799529,0.0004994473,0.00003656135,0.00002191974,0.000007815693,0.00001555061,0.9975389,0.000004514373,0.0006866731,0.00005241691,0.0005089311],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0001881208,0.0001536355,0.9911611,0.0003814656,0.001473189,0.001379923,0.000002216834,0.0008216273,0.0044387],"genre_scores_gemma":[0.4216236,3.607251e-7,0.5773235,0.0002462206,0.0004759069,0.0001097796,0.00005279843,0.00003620909,0.0001315913],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.4214355,"threshold_uncertainty_score":0.9997937,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04957674376864774,"score_gpt":0.2851989741955164,"score_spread":0.2356222304268687,"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."}}