{"id":"W4387586812","doi":"10.2139/ssrn.4600240","title":"Gcsum: Boosting Informativeness and Factual Consistency of Abstractive Summarization with Semantic Graph and Contrastive Learning","year":2023,"lang":"en","type":"preprint","venue":"SSRN Electronic Journal","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Automatic summarization; Boosting (machine learning); Computer science; Natural language processing; Artificial intelligence; Consistency (knowledge bases); Graph; Linguistics; Psychology; Philosophy; Theoretical computer science","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":["research_integrity"],"consensus_categories":[],"category_scores_codex":[0.001356957,0.0002401148,0.0003899182,0.0002745017,0.0002572919,0.0001949751,0.000299643,0.0001373349,5.916286e-7],"category_scores_gemma":[0.0001834485,0.0002085921,0.00004922918,0.0001565021,0.0001175091,0.000494565,0.0003993608,0.002803616,6.30444e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001911164,"about_ca_system_score_gemma":0.001763264,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001701171,"about_ca_topic_score_gemma":0.0002311741,"domain_scores_codex":[0.9978473,0.0001268699,0.0004629309,0.0003429997,0.0003243468,0.0008955432],"domain_scores_gemma":[0.9984216,0.000306613,0.0007648554,0.0001598408,0.00027419,0.00007288187],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0003366264,0.0001231581,0.09338134,0.001270745,0.003125162,0.00007749583,0.04559033,0.1628419,0.0008080137,0.3521055,0.000003754768,0.340336],"study_design_scores_gemma":[0.003020772,0.001323416,0.04252013,0.002669337,0.0003299603,0.001716617,0.02480089,0.366233,0.0005443736,0.5554552,0.00001845052,0.001367867],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3302349,0.0006258755,0.6686503,0.0001060924,0.00008873144,0.0001589166,0.00000175886,0.00004396728,0.00008948956],"genre_scores_gemma":[0.9962975,0.001823284,0.001742273,0.000009409954,0.00003964428,0.000005719337,0.000005574661,0.00001668222,0.00005988861],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.666908,"threshold_uncertainty_score":0.9994969,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01431808181206719,"score_gpt":0.2339334589651947,"score_spread":0.2196153771531275,"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."}}