Summary of Procedural Fairness in Machine Learning, by Ziming Wang et al.
Procedural Fairness in Machine Learning
by Ziming Wang, Changwu Huang, Xin Yao
First submitted to arxiv on: 2 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This research paper proposes novel metrics and methods for ensuring procedural fairness in machine learning models. The authors define individual and group procedural fairness and develop a metric called GPF_{FAE} to evaluate the group procedural fairness of ML models using feature attribution explanation (FAE). They validate the effectiveness of this metric on synthetic and real-world datasets, revealing relationships between procedural and distributive fairness. The paper proposes methods for identifying features that lead to procedural unfairness and improving procedural fairness with minimal impact on model performance. These contributions aim to advance fairness in machine learning by addressing the neglected dimension of procedural fairness. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Machine learning models are not always fair. Researchers have mostly looked at one kind of fairness, called distributive fairness. But there’s another type, called procedural fairness. This paper defines procedural fairness and proposes a new way to measure it using an explainable AI technique called feature attribution explanation (FAE). The authors test this metric on fake and real datasets to see how well it works. They find that procedural and distributive fairness are related. The paper also shows two ways to make ML models more fair by fixing features that cause unfairness. This research can help make machine learning fairer. |
Keywords
* Artificial intelligence * Machine learning
