Summary of Towards Harmless Rawlsian Fairness Regardless Of Demographic Prior, by Xuanqian Wang et al.
Towards Harmless Rawlsian Fairness Regardless of Demographic Prior
by Xuanqian Wang, Jing Li, Ivor W. Tsang, Yew-Soon Ong
First submitted to arxiv on: 4 Nov 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computers and Society (cs.CY); Machine Learning (stat.ML)
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 paper explores the concept of harmless Rawlsian fairness, which aims to achieve fairness in model training without prior knowledge of demographics. The authors propose a simple method called VFair to minimize the variance of training losses and optimize it using a dynamic update approach. They find that regression tasks can significantly improve fairness through VFair, whereas classification tasks do not due to their quantized utility measurements. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making sure AI models are fair without knowing what group they’re from. It proposes a new way to make this happen called VFair. The authors test it and find that some types of problems (regression) can get much more fair using VFair, but others (classification) don’t because of how we measure fairness. |
Keywords
» Artificial intelligence » Classification » Regression
