{"id":"W3173886731","doi":"10.18653/v1/2021.acl-short.119","title":"Demoting the Lead Bias in News Summarization via Alternating Adversarial Learning","year":2021,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Huawei Technologies","keywords":"Automatic summarization; Adversarial system; Computer science; Association (psychology); Joint (building); Natural language processing; Artificial intelligence; Linguistics; Engineering; Philosophy; Epistemology","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":[],"consensus_categories":[],"category_scores_codex":[0.000454315,0.00007411539,0.00008546766,0.00004620651,0.0001390972,0.0001863079,0.0003909656,0.00003655078,0.00002510431],"category_scores_gemma":[0.0003952624,0.00005917852,0.00003313097,0.0003524665,0.000008195419,0.0003409716,0.0003420172,0.0002259586,0.00002148526],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000040771,"about_ca_system_score_gemma":0.0000610318,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008428154,"about_ca_topic_score_gemma":0.0007983485,"domain_scores_codex":[0.9988898,0.0001822714,0.0002382134,0.0002875027,0.0001988652,0.0002033238],"domain_scores_gemma":[0.999382,0.0002062924,0.00007664997,0.000261832,0.00005019449,0.00002308661],"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":[0.000002171749,0.00002669554,0.1159955,0.00001245904,0.00001232018,0.0000447179,0.004123977,0.2983928,0.004255903,0.037778,0.00003566065,0.5393198],"study_design_scores_gemma":[0.0001863885,0.000005174879,0.000782227,0.00001834569,0.000001285472,0.000009835099,0.0002277254,0.9956679,0.001840966,0.0007566441,0.0004285774,0.00007496215],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1050227,0.00002632984,0.8860201,0.001431425,0.0003948831,0.00005061918,2.515502e-8,0.00007049907,0.006983472],"genre_scores_gemma":[0.9248871,0.00000598746,0.07361725,0.0004047081,0.0001799904,0.000002175721,0.000001553945,0.000005875789,0.0008953853],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8198644,"threshold_uncertainty_score":0.241323,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04581204180063356,"score_gpt":0.2636030402067356,"score_spread":0.217790998406102,"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."}}