Desiderata for Explainable AI in statistical production systems of the European Central Bank

Mougan, Carlos; Kanellos, Georgios; Gottron, Thomas

doi:10.48550/arxiv.2107.08045

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To compare disparate treatment between groups we will make use of Caucasian Married as the reference group. It is worth noting that the contribution of the features to the model performance, based on feature importance explanation mechanism [34,37,39,47], is highly relevant. The available data is split into a 50/50 stratified train/test split, maintaining the ratio of each category between train and test set.…”

Section: Compasmentioning

confidence: 99%

Fairness implications of encoding protected categorical attributes

Mougan¹,

Álvarez²,

Patro³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Protected attributes are often presented as categorical features that need to be encoded before feeding them into a machine learning algorithm. Encoding these attributes is paramount as they determine the way the algorithm will learn from the data. Categorical feature encoding has a direct impact on the model performance and fairness. In this work, we compare the accuracy and fairness implications of the two most well-known encoders: one-hot encoding and target encoding. We distinguish between two types of induced bias that can arise while using these encodings and can lead to unfair models. The first type, irreducible bias, is due to direct group category discrimination and a second type, reducible bias, is due to large variance in less statistically represented groups. We take a deeper look into how regularization methods for target encoding can improve the induced bias while encoding categorical features. Furthermore, we tackle the problem of intersectional fairness that arises when mixing two protected categorical features leading to higher cardinality. This practice is a powerful feature engineering technique used for boosting model performance. We study its implications on fairness as it can increase both types of induced bias.

show abstract

Section: Compasmentioning

confidence: 99%