{"id":"W3023058189","doi":"10.1007/978-3-030-47358-7_37","title":"Attending Knowledge Facts with BERT-like Models in Question-Answering: Disappointing Results and Some Explanations","year":2020,"lang":"en","type":"book-chapter","venue":"Lecture notes in computer science","topic":"Topic Modeling","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Sentence; Computer science; Question answering; Natural language processing; Set (abstract data type); Artificial intelligence; Programming language","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0007408703,0.0004659667,0.000480204,0.0007912911,0.0002422869,0.0005478616,0.001588424,0.0002016323,0.000001030284],"category_scores_gemma":[0.0001007515,0.000423196,0.00004871207,0.0005577637,0.0002943402,0.001491687,0.001215577,0.0008242038,0.000006695202],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002616967,"about_ca_system_score_gemma":0.0003807094,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006654532,"about_ca_topic_score_gemma":0.0003858352,"domain_scores_codex":[0.996412,0.0000420478,0.0005980693,0.001778117,0.0006121517,0.000557584],"domain_scores_gemma":[0.9981247,0.0004902232,0.0002390529,0.0008144702,0.0001246219,0.000206916],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001674637,0.00003537202,0.0002136872,0.0001151387,0.00001535232,0.0001808554,0.009095465,0.5139557,0.00008655165,0.2676163,0.00001020951,0.2086586],"study_design_scores_gemma":[0.0003532708,0.00006828153,0.000117412,0.0007847926,0.000004502581,0.00003626834,0.000001147797,0.8851048,0.00007186815,0.1128583,0.0001268132,0.0004725158],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0006371103,0.000366782,0.9935996,0.001871449,0.0006242598,0.0003386269,0.000005843448,0.0001644771,0.002391836],"genre_scores_gemma":[0.6786068,0.00004686047,0.3204507,0.0004997907,0.0002564109,0.00001076739,0.000006616099,0.00003059484,0.00009146253],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6779697,"threshold_uncertainty_score":0.999822,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03398356676494375,"score_gpt":0.2610527462836509,"score_spread":0.2270691795187071,"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."}}