{"id":"W2903495827","doi":"10.1016/j.jss.2020.110542","title":"On testing machine learning programs","year":2020,"lang":"en","type":"preprint","venue":"Journal of Systems and Software","topic":"Software Testing and Debugging Techniques","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":false,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Computer science; Transformative learning; Domain (mathematical analysis); Witness; Software engineering; Field (mathematics); Reliability (semiconductor); Software testing; Artificial intelligence; Software; Data science; Machine learning; Psychology","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":[0.0009253661,0.0002875628,0.0006390171,0.0001961491,0.0001528364,0.0006982565,0.0008109428,0.000206636,4.525336e-7],"category_scores_gemma":[0.003167664,0.0002243303,0.0001379603,0.0002143436,0.0000317115,0.0001385631,0.0007192428,0.001547466,0.000002437921],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006199996,"about_ca_system_score_gemma":0.0001719018,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001435053,"about_ca_topic_score_gemma":5.172283e-7,"domain_scores_codex":[0.9980137,0.0001679556,0.0006917481,0.0003689028,0.0005303037,0.0002274158],"domain_scores_gemma":[0.9970079,0.0007802194,0.001224614,0.0003429784,0.0004311509,0.0002131261],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00006749393,0.000315617,0.1810116,0.00443978,0.0004241014,0.002313008,0.003903515,0.01089078,0.00004563162,0.003614187,0.01435664,0.7786176],"study_design_scores_gemma":[0.002233024,0.01069141,0.008934781,0.04980529,0.0002902208,0.008464132,0.0001096192,0.6414126,0.0000659014,0.2620107,0.01272258,0.003259729],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01789925,0.006267254,0.9675459,0.0004973031,0.00168669,0.0003727702,0.000004863089,0.005600942,0.0001250379],"genre_scores_gemma":[0.7025432,0.00005101614,0.2967846,0.0001016124,0.0004260333,0.00001097009,0.000002905839,0.00003274649,0.00004691419],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7753579,"threshold_uncertainty_score":0.9147924,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06147200893269984,"score_gpt":0.2724679788593244,"score_spread":0.2109959699266246,"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."}}