Dimensionality Reduction Meets Network Science: Sensemaking on UMAP's Internal kNN Graph (arXiv 2507.08728)

Forum topic · 小凯 · 2026-07-13

Summary

A 2025 arXiv paper (2507.08728) by Duen Horng Chau, Donghao Ren, and Fred Hohman argues that UMAP's internally constructed k-nearest-neighbor (kNN) graph is an underused resource for data sensemaking. While typical UMAP workflows focus only on the 2D embedding—which introduces distortion—the internal kNN graph encodes the data manifold in the original high-dimensional space. The authors show that standard network science algorithms applied to this graph yield useful insights: PageRank identifies representative data points, k-core decomposition separates dense core regions from the sparse periphery, and clustering coefficients detect tight-knit neighborhoods of highly similar points. Quantitative and qualitative evaluations on MNIST and Fashion MNIST show these graph-based analyses are competitive with or complementary to purpose-built methods.

Overview

Field: Data Science
Authors: Duen Horng Chau, Donghao Ren, Fred Hohman
Published: 2025-07-12
arXiv: 2507.08728

Abstract

While UMAP is widely used for exploring high-dimensional data, typical workflows focus on its lower-dimensional embedding, largely overlooking the rich k-nearest-neighbor (kNN) graph that UMAP constructs internally. This graph encodes the data manifold in its original high-dimensional space, before the distortion that UMAP's 2D projection introduces.

The authors demonstrate the untapped potential of this internal representation, showing how standard graph algorithms applied to this graph enhance data sensemaking:

1. PageRank identifies representative data points. 2. k-core decomposition reveals dense core regions versus the sparse periphery. 3. Clustering coefficient detects tight-knit neighborhoods with highly similar data points.

Through quantitative and qualitative evaluation on MNIST and Fashion MNIST, they show that these graph-based analyses are competitive with or complementary to purpose-built methods.

Key Takeaway

The kNN graph that UMAP already builds during dimensionality reduction can be reused as a first-class data structure for graph-based analysis—no additional computation pipelines required—enabling richer interpretation of high-dimensional datasets than the 2D embedding alone.

--- *Auto-collected on 2025-07-13*