UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction

Leland McInnes, John Healy, James Melville
2/9/2018
stat.MLcs.CGcs.LG

Official code: lmcinnes/umap (8,281 stars, BSD-3-Clause, last commit 2026-10-16). 1 more implementation below.

Abstract

UMAP (Uniform Manifold Approximation and Projection) is a novel manifold learning technique for dimension reduction. UMAP is constructed from a theoretical framework based in Riemannian geometry and algebraic topology. The result is a practical scalable algorithm that applies to real world data. The UMAP algorithm is competitive with t-SNE for visualization quality, and arguably preserves more of the global structure with superior run time performance. Furthermore, UMAP has no computational restrictions on embedding dimension, making it viable as a general purpose dimension reduction technique for machine learning.

View on arXivView PDF

Code Implementations(2)

lmcinnes/umapOfficial100%
BSD-3-Clause

Uniform Manifold Approximation and Projection

8,281879TeX, ShellJul 2, 20175 days agoBSD-3-Clause
dimensionality-reductionmachine-learningtopological-data-analysisumapvisualization
jlmelville/uwotCommunity100%
GPL-3.0

An R package implementing the UMAP dimensionality reduction method.

37130R, C++Jun 12, 20182 weeks agoGPL-3.0
dimensionality-reductionrumap

Cite this paper

@article{mcinnes2018umap,
  title  = {UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction},
  author = {Leland McInnes and John Healy and James Melville},
  year   = {2018},
  eprint = {1802.03426},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1802.03426}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1802.03426.svg)](https://codeofpaper.com/paper/1802.03426)

Discussion