{"id":"W4200633747","doi":"10.1609/aaai.v36i11.21443","title":"Word Embeddings via Causal Inference: Gender Bias Reducing and Semantic Information Preserving","year":2022,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":24,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"Economic and Social Research Council; Social Sciences and Humanities Research Council of Canada","keywords":"Debiasing; Computer science; Word (group theory); Natural language processing; Artificial intelligence; Inference; Causal inference; Focus (optics); Semantic similarity; Oracle; Cognitive psychology; Psychology; Linguistics; Cognitive science; Mathematics","routes":{"ca_aff":true,"ca_fund":true,"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":[0.0009802709,0.0002453143,0.0002526957,0.0002961706,0.0005983771,0.0005740543,0.002249312,0.00007571279,0.00008116841],"category_scores_gemma":[0.0007522546,0.0002003485,0.00006964697,0.001001248,0.0001729856,0.001748608,0.002118694,0.0006437721,0.00001032801],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009306706,"about_ca_system_score_gemma":0.000124253,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001581172,"about_ca_topic_score_gemma":0.000003596526,"domain_scores_codex":[0.9977598,0.0000384498,0.0006168353,0.0004323217,0.0007748676,0.0003776856],"domain_scores_gemma":[0.9983119,0.0001306609,0.0006051591,0.0003315307,0.0005356406,0.00008509633],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00005903833,0.00008439786,0.00046177,0.0001991899,0.00002154674,0.000001272272,0.01153214,0.0001748367,0.04151571,0.757676,0.0003192,0.1879549],"study_design_scores_gemma":[0.00003542727,0.000179814,0.000167576,0.000205626,0.00001720213,0.00003699186,0.001250737,0.2260236,0.2947118,0.4768503,0.0001354729,0.0003854704],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6143258,0.000458568,0.3671455,0.007054956,0.001209901,0.001694589,0.00001680747,0.001113129,0.0069807],"genre_scores_gemma":[0.9754565,0.00002001349,0.02403617,0.0003056369,0.00003290345,0.00006216031,0.000001165177,0.00001116842,0.00007430601],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3611307,"threshold_uncertainty_score":0.8169976,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07089615856085248,"score_gpt":0.3114886109129865,"score_spread":0.240592452352134,"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."}}