{"id":"W4413364391","doi":"10.1016/j.procs.2025.07.178","title":"Retrieve-Classify-Read: Passage Filtering via Subject Classification for University Question Answering","year":2025,"lang":"en","type":"article","venue":"Procedia Computer Science","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Acadia University","funders":"","keywords":"Computer science; Question answering; Subject (documents); Information retrieval; Artificial intelligence; Natural language processing; World Wide Web","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":[0.003279691,0.00129634,0.001090387,0.002764071,0.0007745314,0.001430718,0.00205278,0.001746606,0.005609877],"category_scores_gemma":[0.01129928,0.0003155906,0.001537921,0.001237861,0.0005712264,0.002406234,0.001461463,0.001540485,0.003940135],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001129268,"about_ca_system_score_gemma":0.00159823,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01390461,"about_ca_topic_score_gemma":0.01265228,"domain_scores_codex":[0.9985708,0.000586733,0.0001070262,0.0004089813,0.0002389161,0.00008757078],"domain_scores_gemma":[0.99537,0.002508917,0.0002399279,0.0008935862,0.0007410605,0.0002465218],"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":[0.001043007,0.0009452947,0.01276717,0.0005790552,0.000296153,0.0003682858,0.0013298,0.02920688,0.03176753,0.003737531,0.02435154,0.8936077],"study_design_scores_gemma":[0.0001157698,0.0008491046,0.008450267,0.00005355728,0.0001916141,0.0003999625,0.0003655134,0.9201342,0.03653907,0.007254431,0.02554679,0.00009969051],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08857667,0.001247738,0.8479666,0.000586505,0.0002845066,0.001178613,0.002866414,0.0541715,0.003121509],"genre_scores_gemma":[0.3811656,0.000526413,0.597157,0.0003814061,0.000250257,0.0006323127,0.01096343,0.00110468,0.007818881],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01390461,"threshold_uncertainty_score":0.02764732,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02297700179631848,"score_gpt":0.2571357960003268,"score_spread":0.2341587942040083,"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."}}