{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004543242,0.0001349185,0.000232041,0.000162211,0.0002253381,0.00004688001,0.001996732,0.00003077366,0.00003807919],"category_scores_gemma":[0.0001818935,0.0001055536,0.00007173675,0.0009142236,0.00005209803,0.0004925868,0.001817501,0.0002400897,0.000006159474],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001647175,"about_ca_system_score_gemma":0.00008931754,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008663544,"about_ca_topic_score_gemma":0.0003187466,"domain_scores_codex":[0.9975075,0.0006998782,0.0004261557,0.0005917319,0.0003868328,0.0003879548],"domain_scores_gemma":[0.9978815,0.0006415298,0.00008915761,0.001211619,0.0001077428,0.00006844912],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001043116,0.0001055081,0.00003038745,0.00001963145,0.00000439324,0.000003540297,0.01139933,0.03897607,0.002719795,0.5505528,0.00002001351,0.3961581],"study_design_scores_gemma":[0.00007298311,0.0002042239,0.0001758711,0.000005986524,0.000002405456,0.000005206525,0.02516993,0.5360088,0.08307361,0.3526701,0.002374933,0.0002358971],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01239313,0.0003469907,0.9704324,0.005091992,0.0002903088,0.000873155,0.000001180764,0.00008485353,0.01048597],"genre_scores_gemma":[0.8908263,0.00001704026,0.1077338,0.0002623022,0.00001503964,0.001019063,1.748779e-7,0.000008612035,0.0001175961],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8784332,"threshold_uncertainty_score":0.4304353,"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."}}