{"id":"W3098123193","doi":"10.18653/v1/2020.codi-1.13","title":"Do We Really Need That Many Parameters In Transformer For Extractive Summarization? Discourse Can Help !","year":2020,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Huawei Technologies","keywords":"Automatic summarization; Transformer; Computer science; Artificial intelligence; Natural language processing; Sentence; Language model; Machine learning","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":[0.005102359,0.001960995,0.001579007,0.001346092,0.001429069,0.003867158,0.002306296,0.002683926,0.02281682],"category_scores_gemma":[0.02952355,0.001044067,0.00112474,0.001250374,0.00196873,0.02774908,0.003316741,0.005272944,0.01461758],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001180853,"about_ca_system_score_gemma":0.001492767,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002872959,"about_ca_topic_score_gemma":0.006120008,"domain_scores_codex":[0.9975781,0.001201869,0.0001538294,0.0005519232,0.0003295259,0.0001846794],"domain_scores_gemma":[0.98976,0.005664206,0.000440691,0.002485647,0.001210183,0.0004392764],"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":[0.001286442,0.0002271557,0.002884377,0.001314028,0.0002323346,0.0003179995,0.003504308,0.009126462,0.0383521,0.09918614,0.05099892,0.7925698],"study_design_scores_gemma":[0.000309018,0.0005044207,0.002299594,0.0008842827,0.0005899805,0.0008577961,0.003782762,0.1310708,0.05400408,0.5526837,0.2526455,0.0003682244],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01689079,0.004377115,0.925132,0.02134384,0.001010455,0.0001899068,0.0008115824,0.01673277,0.01351143],"genre_scores_gemma":[0.3595204,0.003214387,0.6069119,0.005511379,0.001110618,0.0003080903,0.00189209,0.006916904,0.01461421],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02281682,"threshold_uncertainty_score":0.07632983,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04987185070598387,"score_gpt":0.2862067534909244,"score_spread":0.2363349027849405,"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."}}