Research
Some Theoretical Limitations of t-SNE
arXiv:2604.13295v1 Announce Type: new Abstract: t-SNE has gained popularity as a dimension reduction technique, especially for visualizing data. It is well-known that all dimension reduction technique
arXiv:2604.13295v1 Announce Type: new Abstract: t-SNE has gained popularity as a dimension reduction technique, especially for visualizing data. It is well-known that all dimension reduction techniques may lose important features of the data. We provide a mathematical framework for understanding this loss for t-SNE by establishing a number of results in different scenarios showing how important features of data are lost by using t-SNE.
Related
- Bi-Lipschitz Autoencoder With Injectivity Guarantee
- The Linear Centroids Hypothesis: How Deep Network Features Represent Data
- Scalable unsupervised feature selection via weight stability
- UniPROT: Uniform Prototype Selection via Partial Optimal Transport with Submodular Guarantees
Source: arXiv cs.LG | 2026-04-16