{"id":"W3208730084","doi":"10.48550/arxiv.2110.14081","title":"A Controlled Experiment of Different Code Representations for Learning-Based Bug Repair","year":2021,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Software Engineering Research","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Code (set theory); Source code; Representation (politics); Perspective (graphical); Point (geometry); Process (computing); Abstract syntax; Programming language; Syntax; Artificial intelligence; Set (abstract data type); Mathematics","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":[0.00384535,0.001513572,0.0006551414,0.000518261,0.0003454323,0.0008082222,0.001830836,0.001995902,0.00387854],"category_scores_gemma":[0.02724168,0.0005229925,0.0008096992,0.0003878183,0.001055841,0.002208356,0.001185398,0.002895957,0.001045923],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006143589,"about_ca_system_score_gemma":0.0009710259,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001706843,"about_ca_topic_score_gemma":0.002255333,"domain_scores_codex":[0.99765,0.0008666157,0.0002400236,0.0007661683,0.0002868649,0.0001903969],"domain_scores_gemma":[0.9666806,0.02612286,0.001488402,0.003351906,0.001520594,0.0008356955],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.03743224,0.04696339,0.04488704,0.005913874,0.001055911,0.001323561,0.004898876,0.281568,0.130258,0.008107062,0.02269331,0.4148988],"study_design_scores_gemma":[0.009237399,0.06228633,0.03045085,0.0007513195,0.001043849,0.0006383788,0.001595481,0.7228447,0.1312103,0.01415264,0.02539308,0.0003958155],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9731006,0.0004394085,0.02001852,0.0002822322,0.000238731,0.001210164,0.001104172,0.001942225,0.001663946],"genre_scores_gemma":[0.9461356,0.0002873737,0.04503523,0.0004490294,0.00006805542,0.002519342,0.00204222,0.0003087106,0.003154441],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.00387854,"threshold_uncertainty_score":0.02033639,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07040668802292974,"score_gpt":0.2390651119295893,"score_spread":0.1686584239066596,"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."}}