Summary of Fairness-aware Meta-learning Via Nash Bargaining, by Yi Zeng et al.
Fairness-Aware Meta-Learning via Nash Bargaining
by Yi Zeng, Xuelin Yang, Li Chen, Cristian Canton Ferrer, Ming Jin, Michael I. Jordan, Ruoxi Jia
First submitted to arxiv on: 11 Jun 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 paper addresses issues of group-level fairness in machine learning by adjusting model parameters based on specific fairness objectives over a sensitive-attributed validation set. The authors introduce a two-stage meta-learning framework to resolve hypergradient conflicts, which can cause unstable convergence and compromise model performance and fairness. In the first stage, they use the Nash Bargaining Solution (NBS) to steer the model toward the Pareto front, while in the second stage, they optimize with respect to specific fairness goals. The method is supported by theoretical results, including proofs of NBS for gradient aggregation free from linear independence assumptions, Pareto improvement, and monotonic improvement in validation loss. Empirical effects are shown across various fairness objectives in six key fairness datasets and two image classification tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps make sure machine learning models are fair to different groups of people. Right now, some models can be biased against certain groups. The authors came up with a new way to fix this problem by using a special kind of math called meta-learning. They created a two-step process: first, they used a special formula to make the model more balanced and fair, and then they fine-tuned it to fit specific goals for fairness. This method works well in many different situations and can help make sure machine learning models are fairer. |
Keywords
» Artificial intelligence » Image classification » Machine learning » Meta learning
