Summary of Towards Better Spherical Sliced-wasserstein Distance Learning with Data-adaptive Discriminative Projection Direction, by Hongliang Zhang et al.
Towards Better Spherical Sliced-Wasserstein Distance Learning with Data-Adaptive Discriminative Projection Direction
by Hongliang Zhang, Shuo Chen, Lei Luo, Jian Yang
First submitted to arxiv on: 26 Dec 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 The proposed Discriminative Spherical Sliced-Wasserstein (DSSW) distance improves upon the original SSW by incorporating projection direction weights. This is achieved through two types of projected energy functions: a non-parametric deterministic function that adapts to different data distributions, and a neural network-induced function that learns optimal weights from data projections. The DSSW distance is evaluated across various machine learning tasks, including gradient flows, density estimation on real earth data, and self-supervised learning, demonstrating improved performance compared to state-of-the-art methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new way to measure how similar or different two groups of spherical data are. This method, called Discriminative Spherical Sliced-Wasserstein (DSSW), is better than the original method because it takes into account which directions in space are most important for comparing the two groups. The DSSW method uses two types of calculations to determine these important directions: one that works well for many different data sets and another that learns from the data itself. The paper shows that this new method performs better than other methods on various tasks, such as predicting how objects move or estimating the distribution of real-world data. |
Keywords
» Artificial intelligence » Density estimation » Machine learning » Neural network » Self supervised
