English

Some Theoretical Limitations of t-SNE

Machine Learning 2026-04-16 v1 Probability Machine Learning

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.

Keywords

Cite

@article{arxiv.2604.13295,
  title  = {Some Theoretical Limitations of t-SNE},
  author = {Rupert Li and Elchanan Mossel},
  journal= {arXiv preprint arXiv:2604.13295},
  year   = {2026}
}

Comments

19 pages, 7 figures