{"id":"W7106833217","doi":"10.48448/4qjf-h498","title":"Topic-Guided Reinforcement Learning with LLMs for Enhancing Multi-Document Summarization","year":2025,"lang":"","type":"other","venue":"Open MIND","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Automatic summarization; Reinforcement learning; Selection (genetic algorithm); Coherence (philosophical gambling strategy); Key (lock); Reinforcement","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":[0.001444618,0.0008720427,0.0011752,0.0007832801,0.0003877872,0.0008311116,0.001074364,0.0009795388,0.001776392],"category_scores_gemma":[0.005201227,0.000305179,0.0004790264,0.0007643517,0.000435006,0.001609658,0.0009830071,0.001228895,0.0008822663],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006988408,"about_ca_system_score_gemma":0.0009944319,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002713055,"about_ca_topic_score_gemma":0.004169603,"domain_scores_codex":[0.9991929,0.0003168817,0.00004725405,0.0002258352,0.0001565609,0.0000606422],"domain_scores_gemma":[0.9981121,0.001117648,0.0002149328,0.0001802383,0.0002783408,0.00009686956],"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":[0.0003989651,0.0003490345,0.001943772,0.000325659,0.0001370522,0.0002017777,0.0004428608,0.4215018,0.02571659,0.007062595,0.009478051,0.5324418],"study_design_scores_gemma":[0.00002567163,0.00006253648,0.0001712475,0.000009095332,0.00001604487,0.00001746264,0.00001955992,0.9910431,0.003342007,0.003920959,0.001363398,0.000008873722],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0316541,0.001355227,0.9607227,0.0004913817,0.0001089123,0.0000774784,0.0001693779,0.004057187,0.001363619],"genre_scores_gemma":[0.6885052,0.0005260821,0.3030395,0.0004980864,0.0003221917,0.0002054618,0.0009363936,0.0005018474,0.005465223],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.002713055,"threshold_uncertainty_score":0.007640004,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04165672396215413,"score_gpt":0.3418393731655975,"score_spread":0.3001826492034434,"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."}}