{"id":"W4407580437","doi":"10.1145/3717061","title":"An Empirical Study of Retrieval-Augmented Code Generation: Challenges and Opportunities","year":2025,"lang":"en","type":"article","venue":"ACM Transactions on Software Engineering and Methodology","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Code (set theory); Data science; Information retrieval; 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":[],"consensus_categories":[],"category_scores_codex":[0.0005517748,0.0001129372,0.0002119972,0.0002583262,0.00007369921,0.00002239434,0.0002541061,0.0000969151,0.000001347303],"category_scores_gemma":[0.0002703198,0.0001049759,0.00001704894,0.0001350594,0.00003345165,0.0001552985,0.00002089966,0.0001742833,4.741695e-8],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001242092,"about_ca_system_score_gemma":0.00002584977,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000006335299,"about_ca_topic_score_gemma":0.000007242664,"domain_scores_codex":[0.9991345,0.000226591,0.0001702794,0.0002788531,0.00008037564,0.0001093552],"domain_scores_gemma":[0.9988062,0.0006480544,0.0000321108,0.0004004761,0.00006610344,0.00004699922],"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":[0.00005711172,0.000364255,0.0000984836,0.0001947765,0.0001473207,0.00002019513,0.007268498,0.00113869,0.007002832,0.005540127,0.00002146465,0.9781463],"study_design_scores_gemma":[0.0111674,0.02105138,0.02155278,0.001258332,0.001202714,0.000845546,0.01237919,0.2612433,0.5974359,0.05855048,0.008918506,0.004394522],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06325217,0.005237664,0.9303295,0.0005377175,0.0001634789,0.0001038044,0.0000023593,0.0003714798,0.000001802576],"genre_scores_gemma":[0.2757795,0.0007651281,0.7233033,0.00009121055,0.0000110916,0.0000137281,8.513636e-7,0.000005741418,0.00002946964],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9737517,"threshold_uncertainty_score":0.4280794,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.209148202889168,"score_gpt":0.3915554796441723,"score_spread":0.1824072767550042,"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."}}