{"id":"W3174481471","doi":"10.18653/v1/2021.acl-long.564","title":"Mind Your Outliers! Investigating the Negative Impact of Outliers on Active Learning for Visual Question Answering","year":2021,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":50,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Open Philanthropy Project; Canadian Institute for Advanced Research","keywords":"Question answering; Outlier; Computational linguistics; Natural language processing; Computer science; Artificial intelligence; Corpus linguistics; Volume (thermodynamics); Linguistics; Information retrieval; Philosophy","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004210447,0.0001644166,0.0001838076,0.000082438,0.0003451342,0.0001174473,0.0003765931,0.00006024725,0.00001571343],"category_scores_gemma":[0.001734133,0.000120981,0.0001488102,0.0004727411,0.00008234368,0.0002511607,0.0001491289,0.0003842169,0.00001259373],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001353361,"about_ca_system_score_gemma":0.000205599,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006452985,"about_ca_topic_score_gemma":0.0000227414,"domain_scores_codex":[0.9986174,0.0002386744,0.0002471133,0.0004103948,0.0002316401,0.0002547748],"domain_scores_gemma":[0.998129,0.0009485353,0.0002527672,0.0003291086,0.0002483874,0.00009221854],"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":[0.00005814356,0.0002557319,0.03744756,0.00004004092,0.0002965634,0.000003052123,0.03326857,0.5119997,0.07309273,0.03069483,0.00009391431,0.3127491],"study_design_scores_gemma":[0.0004167564,0.0003359437,0.1153609,0.0000513625,0.00001467789,0.000004589635,0.001553915,0.8425766,0.03792261,0.001533884,0.00004767815,0.0001811691],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6809281,0.000005898275,0.3164752,0.001022831,0.00005334292,0.0002746615,0.000002628532,0.00007320497,0.001164137],"genre_scores_gemma":[0.9361259,9.819245e-7,0.06345762,0.00008924067,0.00005199002,0.0000531887,0.000008291338,0.00001521493,0.0001975858],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3305768,"threshold_uncertainty_score":0.4933464,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02975810514562487,"score_gpt":0.3674712042243422,"score_spread":0.3377130990787173,"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."}}