{"id":"W2963723885","doi":"","title":"Understanding the Origins of Bias in Word Embeddings","year":2019,"lang":"en","type":"article","venue":"International Conference on Machine Learning","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":58,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Word (group theory); Word embedding; Natural language processing; Embedding; Artificial intelligence; TRACE (psycholinguistics); Harm; Machine learning; Linguistics; Psychology","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":[0.003172817,0.0005175489,0.000473298,0.0016221,0.000544564,0.001759768,0.0005250723,0.0008547837,0.0009082171],"category_scores_gemma":[0.04818755,0.0004770156,0.0003598117,0.001608691,0.001291101,0.005480413,0.001558029,0.001543098,0.0005293573],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009233513,"about_ca_system_score_gemma":0.001099742,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004065749,"about_ca_topic_score_gemma":0.005401797,"domain_scores_codex":[0.9978573,0.0009006187,0.0001574326,0.0004023217,0.0005519471,0.0001304162],"domain_scores_gemma":[0.9840721,0.009261594,0.00166935,0.002558984,0.00229505,0.0001428205],"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":[0.0003789664,0.0001376245,0.09534712,0.000577245,0.000242481,0.0004587252,0.004716972,0.2014243,0.03820248,0.1334704,0.006925682,0.5181179],"study_design_scores_gemma":[0.00003288677,0.00007948787,0.01961598,0.0001557632,0.0000571231,0.0004067018,0.0009354815,0.7184461,0.03037999,0.2163332,0.01347788,0.00007950432],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3268034,0.001552564,0.6644776,0.001499221,0.0001608544,0.00006812342,0.0004973945,0.0007151815,0.004225702],"genre_scores_gemma":[0.8819226,0.001012119,0.1139049,0.000243924,0.00009984298,0.0001076224,0.000698898,0.0003318803,0.001677995],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004065749,"threshold_uncertainty_score":0.01677966,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1021607067859051,"score_gpt":0.341356417713073,"score_spread":0.2391957109271678,"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."}}