Summary of Towards Marginal Fairness Sliced Wasserstein Barycenter, by Khai Nguyen and Hai Nguyen and Nhat Ho
Towards Marginal Fairness Sliced Wasserstein Barycenter
by Khai Nguyen, Hai Nguyen, Nhat Ho
First submitted to arxiv on: 13 May 2024
Categories
- Main: Machine Learning (stat.ML)
- Secondary: Graphics (cs.GR); Machine Learning (cs.LG)
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 paper proposes a novel method for achieving marginal fairness in the sliced Wasserstein barycenter (SWB) by defining a constrained SWB problem, which they call the marginal fairness sliced Wasserstein barycenter (MFSWB). The authors recognize that the uniform weighted SWB is not optimal for achieving marginal fairness due to the heterogeneous structure of marginals. They propose two hyperparameter-free and computationally tractable surrogate MFSWB problems that implicitly minimize distances to marginals and encourage marginal fairness. To improve efficiency, they select a slicing distribution and introduce a new slicing distribution that focuses on marginally unfair projecting directions. The paper discusses the relationships between the proposed problems and the sliced multi-marginal Wasserstein distance. Experimental results demonstrate the favorable performance of the proposed surrogate MFSWB problems in applications such as 3D point-clouds averaging, color harmonization, and training of sliced Wasserstein autoencoder with class-fairness representation. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper is about finding a way to make sure that when we average things together, it’s fair for everyone. They’re trying to fix a problem in a method called the sliced Wasserstein barycenter (SWB) that makes sure everything averages out equally. The authors realize that this method isn’t good enough because it doesn’t take into account how different groups might be treated unfairly. So they came up with new ways to make sure everyone is treated fairly, and tested them on things like averaging 3D point clouds together or making colors look nice. |
Keywords
» Artificial intelligence » Autoencoder » Hyperparameter
