{"id":"W4378942602","doi":"10.1145/3597926.3598135","title":"How Effective Are Neural Networks for Fixing Security Vulnerabilities","year":2023,"lang":"en","type":"preprint","venue":"","topic":"Software Testing and Debugging Techniques","field":"Computer Science","cited_by":90,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta; York University","funders":"National Science Foundation","keywords":"Computer science; Task (project management); Vulnerability (computing); Code (set theory); Software security assurance; Secure coding; Computer security; Source code; Software; Artificial neural network; Software bug; Automation; Programming language; Software engineering; Artificial intelligence; Information security; Engineering; Security service","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003086173,0.001406625,0.0007414048,0.001078236,0.000422634,0.002305769,0.0009718284,0.003234642,0.003046815],"category_scores_gemma":[0.02958437,0.0005349159,0.0004321634,0.0005418243,0.00108227,0.008299098,0.000807499,0.002060202,0.001347335],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009698254,"about_ca_system_score_gemma":0.001147727,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005457677,"about_ca_topic_score_gemma":0.006748009,"domain_scores_codex":[0.9981365,0.0005960696,0.0000737486,0.0005323411,0.0003951627,0.0002661184],"domain_scores_gemma":[0.991687,0.005540934,0.0006867127,0.0009766413,0.0007986865,0.0003100418],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001035269,0.0005142996,0.021983,0.0005151174,0.0005800984,0.0001390658,0.0001827981,0.2689991,0.01279367,0.01418246,0.02180167,0.6572735],"study_design_scores_gemma":[0.00008922849,0.0003129583,0.007123369,0.0001959837,0.0002423819,0.0001637897,0.0002255493,0.9087669,0.0148015,0.06278399,0.00522646,0.00006792458],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5140172,0.02228146,0.3751123,0.0331829,0.001755323,0.0001646163,0.001619797,0.007073826,0.0447926],"genre_scores_gemma":[0.9543778,0.002882225,0.03738718,0.0009998658,0.0002389659,0.00004452619,0.0004757416,0.0002639316,0.003329819],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.005457677,"threshold_uncertainty_score":0.01632148,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03806044719484671,"score_gpt":0.2861846731068167,"score_spread":0.24812422591197,"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."}}