Summary of Point Cloud Classification Via Deep Set Linearized Optimal Transport, by Scott Mahan et al.
Point Cloud Classification via Deep Set Linearized Optimal Transport
by Scott Mahan, Caroline Moosmüller, Alexander Cloninger
First submitted to arxiv on: 2 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Machine Learning (stat.ML)
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 Deep Set Linearized Optimal Transport (DSLOT) is an efficient algorithm for simultaneously embedding point clouds into an L2-space, preserving low-dimensional structures within the Wasserstein space. DSLOT constructs a classifier to distinguish between classes of point clouds by leveraging the observation that L2-distances between optimal transport maps approximate Wasserstein-2 distances under certain assumptions. The approach employs input convex neural networks (ICNNs) and trains a discriminator network to create a permutation-invariant classifier. Our algorithm outperforms the standard deep set method on a flow cytometry dataset with limited labeled point clouds. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary We introduce an algorithm that helps computers understand how similar or different groups of tiny points are in space. It does this by taking these points and putting them into a special kind of math space called L2-space. The algorithm is good at keeping the important patterns within this space, which lets it tell the difference between groups of points that belong to different categories. This is useful for things like sorting cells in biology or recognizing pictures in computer vision. Our method uses special types of artificial intelligence called neural networks and works well even when we don’t have many examples to learn from. |
Keywords
* Artificial intelligence * Embedding
