Summary of Fairly Accurate: Optimizing Accuracy Parity in Fair Target-group Detection, by Soumyajit Gupta et al.
Fairly Accurate: Optimizing Accuracy Parity in Fair Target-Group Detection
by Soumyajit Gupta, Venelin Kovatchev, Maria De-Arteaga, Matthew Lease
First submitted to arxiv on: 16 Jul 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 A novel approach is proposed for identifying the demographic group targeted by a social media post, also known as target detection, while ensuring fairness and equal protection to all groups. The authors introduce Accuracy Parity (AP) as their fairness objective, which aims to balance detection accuracy across groups. To achieve this, they develop Group Accuracy Parity (GAP), a differentiable loss function that maps one-to-one with AP. Experiments demonstrate that GAP effectively mitigates bias and improves upon traditional loss functions. |
Low | GrooveSquid.com (original content) | Low Difficulty Summary To detect the demographic group targeted by a social media post, researchers propose a new approach that balances detection accuracy across groups to ensure fairness. They aim to provide equal protection to all groups, introducing Accuracy Parity (AP) as their fairness objective. To achieve this, they develop Group Accuracy Parity (GAP), a differentiable loss function. The study shows that GAP is more effective in reducing bias than other commonly used loss functions. |
Keywords
» Artificial intelligence » Loss function