A survey of bias in Machine Learning through the prism of Statistical Parity for the Adult Data Set

2020/03/31 by Philippe Besse, Eustasio del Barrio, Besse, Philippe +8 · 10 citations
Computer Science · Mathematics · Social Sciences · #Advanced Causal Inference Techniques #Artificial intelligence #Binary classification #Binary number #Computer science #Computers and Society (cs.CY) #FOS: Computer and information sciences #Income, Poverty, and Inequality #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine learning #Mathematics #Population #Set (abstract data type) #Statistical Methods and Inference #Support vector machine #Task (project management) #cs.CY #cs.LG #stat.ML

paper · pdf · doi:10.48550/arxiv.2003.14263

published in arXiv (Cornell University) (Cornell University)

openalex publication_date 2020/03/31 · arxiv created 2020/04/06 · arxiv updated 2020/04/07 · openalex created_date 2020/04/10 · openalex updated_date 2026/07/28

Abstract

Applications based on Machine Learning models have now become an indispensable part of the everyday life and the professional world. A critical question then recently arised among the population: Do algorithmic decisions convey any type of discrimination against specific groups of population or minorities? In this paper, we show the importance of understanding how a bias can be introduced into automatic decisions. We first present a mathematical framework for the fair learning problem, specifically in the binary classification setting. We then propose to quantify the presence of bias by using the standard Disparate Impact index on the real and well-known Adult income data set. Finally, we check the performance of different approaches aiming to reduce the bias in binary classification outcomes. Importantly, we show that some intuitive methods are ineffective. This sheds light on the fact trying to make fair machine learning models may be a particularly challenging task, in particular when the training observations contain a bias.

Citations

• Equality of Opportunity in Supervised Learning
• Fairness in Criminal Justice Risk Assessments: The State of the Art
• Counterfactual Fairness
• Calibration for the (Computationally-Identifiable) Masses
• Obtaining fairness using optimal transport theory

Cited by

• Measuring What Matters: Connecting AI Ethics Evaluations to System Attributes, Hazards, and Harms
• OT-FairBoost: Optimal Transport-Guided Gradient Boosting for Fairness Regularization on Tabular Data
• Machine Learning with Multitype Protected Attributes: Intersectional Fairness through Regularisation
• The World According to LLMs: How Geographic Origin Influences LLMs' Entity Deduction Capabilities
• Exposing the Illusion of Fairness: Auditing Vulnerabilities to Distributional Manipulation Attacks
• Fairness-Aware Grouping for Continuous Sensitive Variables: Application for Debiasing Face Analysis with respect to Skin Tone
• Size-adaptive Hypothesis Testing for Fairness
• When majority rules, minority loses: bias amplification of gradient descent
• End-To-End Bias Mitigation: Removing Gender Bias in Deep Learning
• Fairness is in the details: Face Dataset Auditing

Related

• Coping with Mistreatment in Fair Algorithms
• Social Bias Meets Data Bias: The Impacts of Labeling and Measurement Errors on Fairness Criteria
• Evaluating Fairness Metrics in the Presence of Dataset Bias
• Two-stage Algorithm for Fairness-aware Machine Learning
• Dissecting Causal Biases