{"id":"W4287887550","doi":"10.18653/v1/2022.trustnlp-1.8","title":"Challenges in Applying Explainability Methods to Improve the Fairness of NLP Models","year":2022,"lang":"en","type":"article","venue":"","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":38,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Computer science; Artificial intelligence; Machine learning","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":[0.1546525,0.001984169,0.003076294,0.00443374,0.003436678,0.01154139,0.005839009,0.003791785,0.004549833],"category_scores_gemma":[0.4260271,0.0013908,0.003213984,0.003440902,0.01101514,0.02056496,0.01150793,0.01312199,0.0009533032],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005742216,"about_ca_system_score_gemma":0.01030221,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006116166,"about_ca_topic_score_gemma":0.005474397,"domain_scores_codex":[0.8684296,0.09767743,0.006050861,0.009121595,0.01721461,0.001505858],"domain_scores_gemma":[0.4623174,0.4631885,0.0122358,0.04468704,0.01572732,0.001843949],"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":[0.0001537814,0.0001254414,0.007600085,0.001044124,0.000735809,0.0001507426,0.002449567,0.04185709,0.0007730902,0.796532,0.005049626,0.1435286],"study_design_scores_gemma":[0.00003258201,0.00003004769,0.0004670009,0.0002698693,0.00004916843,0.00004300772,0.0001905126,0.06931078,0.0006096823,0.9222873,0.006675959,0.00003407113],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.005318585,0.002626652,0.9719524,0.01455904,0.0003529777,0.0001933631,0.0002090572,0.000452869,0.004335065],"genre_scores_gemma":[0.2953225,0.003001934,0.6891941,0.005806585,0.001759574,0.001136643,0.0006225212,0.0009698494,0.002186225],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.1546525,"threshold_uncertainty_score":0.8178902,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1524628603607702,"score_gpt":0.3699296182615067,"score_spread":0.2174667579007365,"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."}}