{"id":"W2155056246","doi":"10.1093/comjnl/bxs018","title":"A Source Code Similarity System for Plagiarism Detection","year":2012,"lang":"en","type":"article","venue":"The Computer Journal","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":109,"is_retracted":false,"has_abstract":true,"ca_institutions":"Athabasca University","funders":"","keywords":"Library science; Computer science; Similarity (geometry); Information retrieval; Source code; World Wide Web; Database; Operating system; Artificial intelligence","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":["sts"],"consensus_categories":[],"category_scores_codex":[0.006118626,0.00008848985,0.000126659,0.00004160528,0.002070225,0.0001571902,0.0004102248,0.0004162097,0.00002219652],"category_scores_gemma":[0.0001102482,0.0000607667,0.0001201236,0.00009306213,0.0001143268,0.000232105,0.00004493108,0.0017157,0.00003468674],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001836982,"about_ca_system_score_gemma":0.00007123258,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001503071,"about_ca_topic_score_gemma":0.00006730249,"domain_scores_codex":[0.9983636,0.0005795202,0.0002103426,0.00007957988,0.0003233041,0.00044371],"domain_scores_gemma":[0.9989201,0.0005134419,0.0001636949,0.00009522343,0.0001323165,0.0001751529],"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":[0.0002886022,0.0001879853,0.006805251,0.0001199053,0.0004072468,0.000006816102,0.2242665,0.001082013,0.0006143552,0.1677851,0.1329468,0.4654895],"study_design_scores_gemma":[0.0005418913,0.00006857843,0.001128621,0.0000753206,0.00008889884,0.0004494315,0.006299549,0.01729472,0.0006097971,0.005079003,0.9681095,0.0002547132],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03398203,0.0002409163,0.955578,0.003316255,0.004116686,0.0002165539,0.000004867695,0.00008440032,0.002460286],"genre_scores_gemma":[0.9860751,0.00002672533,0.002693725,0.0007274063,0.01002367,0.000005364284,6.528538e-7,0.000009273996,0.000438097],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9528843,"threshold_uncertainty_score":0.999229,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03357069603471791,"score_gpt":0.294971855487683,"score_spread":0.2614011594529651,"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."}}