Summary of Fairness Without Harm: An Influence-guided Active Sampling Approach, by Jinlong Pang et al.
Fairness Without Harm: An Influence-Guided Active Sampling Approach
by Jinlong Pang, Jialu Wang, Zhaowei Zhu, Yuanshun Yao, Chen Qian, Yang Liu
First submitted to arxiv on: 20 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 The proposed method aims to train machine learning models that mitigate group fairness disparity without compromising model accuracy. By acquiring more data, the algorithm scores each new example based on its influence on fairness and accuracy, and then selects a certain number of examples for training. This approach does not rely on sensitive attribute annotations, making it suitable for real-world applications where privacy and safety concerns are paramount. Theoretical analysis shows that acquiring more data can improve fairness without causing harm, while extensive experiments on real-world data demonstrate the effectiveness of the proposed algorithm. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Machine learning models should be fair to all groups. A big problem is when a model is good for some groups but not others. This paper wants to solve this problem by making sure the model is fair and accurate at the same time. They found that having more data helps, as long as you get the right kind of data. The method they propose doesn’t need special labels for sensitive information, which is important because we don’t want to reveal private details. This approach was tested on real-world data and worked well. |
Keywords
* Artificial intelligence * Machine learning
