mlsanityData to 14 Oct 2026

Stop Misusing t-SNE and UMAP for Visual Analytics

arXiv 2506.08725 ↗First posted 10 Jun 2025cs.HC cs.LGcited by 8

Not on the map. Why?

Abstract (arXiv)

Misuses of t-SNE and UMAP in visual analytics have become increasingly common. For example, although t-SNE and UMAP projections often do not faithfully reflect the original distances between clusters, practitioners frequently use them to investigate inter-cluster relationships. We investigate why this misuse occurs, and discuss methods to prevent it. To that end, we first review 136 papers to verify the prevalence of the misuse. We then interview researchers who have used dimensionality reduction (DR) to understand why such misuse occurs. Finally, we interview DR experts to examine why previous efforts failed to address the misuse. We find that the misuse of t-SNE and UMAP stems primarily from limited DR literacy among practitioners, and that existing attempts to address this issue -- mostly based on academic papers -- have been ineffective. Based on these insights, we discuss potential future research directions to mitigate the misuse.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Builds on 39 of its 105 references that are in the map. First posted on arXiv 10 Jun 2025.

Builds on

39 of 105 references.

References · 39

  1. 2018UMAP: Uniform Manifold Approximation and Projection for Dimension Reductioncited by 2,898
  2. 2023Siren's Song in the AI Ocean: A Survey on Hallucination in Large Language Modelscited by 698
  3. 2020Understanding How Dimension Reduction Tools Work: An Empirical Approach to Deciphering t-SNE, UMAP, TriMAP, and PaCMAP for Data Visualizationcited by 173
  4. 2019Topological Autoencoderscited by 127
  5. 2020Parametric UMAP embeddings for representation and semi-supervised learningcited by 104
  6. 2019TriMap: Large-scale Dimensionality Reduction Using Tripletscited by 60
  7. 2017ActiVis: Visual Exploration of Industry-Scale Deep Neural Network Modelscited by 56
  8. 2014A survey of dimensionality reduction techniquescited by 50
  9. 2018An Analysis of the t-SNE Algorithm for Data Visualizationcited by 49
  10. 2024VisEval: A Benchmark for Data Visualization in the Era of Large Language Modelscited by 45
  11. 2020t-viSNE: Interactive Assessment and Interpretation of t-SNE Projectionscited by 43
  12. 2018GAN Lab: Understanding Complex Deep Generative Models using Interactive Visual Experimentationcited by 41
27 more
  1. 2023How Does Attention Work in Vision Transformers? A Visual Analytics Attemptcited by 30
  2. 2021Revisiting Dimensionality Reduction Techniques for Visual Cluster Analysis: An Empirical Studycited by 27
  3. 2021Measuring and Explaining the Inter-Cluster Reliability of Multidimensional Projectionscited by 20
  4. 2025Unveiling High-dimensional Backstage: A Survey for Reliable Visual Analytics with Dimensionality Reductioncited by 16
  5. 2023ZADU: A Python Library for Evaluating the Reliability of Dimensionality Reduction Embeddingscited by 15
  6. 2021Interactive Dimensionality Reduction for Comparative Analysiscited by 12
  7. 2022Uniform Manifold Approximation with Two-phase Optimizationcited by 12
  8. 2021VAC-CNN: A Visual Analytics System for Comparative Studies of Deep Convolutional Neural Networkscited by 11
  9. 2023Classes are not Clusters: Improving Label-based Evaluation of Dimensionality Reductioncited by 11
  10. 2020A Visual Analytics Framework for Reviewing Multivariate Time-Series Data with Dimensionality Reductioncited by 9
  11. 2024Leveraging Foundation Models for Crafting Narrative Visualization: A Surveycited by 9
  12. 2022Feature Learning for Nonlinear Dimensionality Reduction toward Maximal Extraction of Hidden Patternscited by 8
  13. 2023Large-Scale Evaluation of Topic Models and Dimensionality Reduction Methods for 2D Text Spatializationcited by 8
  14. 2023CLAMS: A Cluster Ambiguity Measure for Estimating Perceptual Variability in Visual Clusteringcited by 8
  15. 2023Class-constrained t-SNE: Combining Data Features and Class Probabilitiescited by 8
  16. 2025A Critical Analysis of the Usage of Dimensionality Reduction in Four Domainscited by 7
  17. 2022HardVis: Visual Analytics to Handle Instance Hardness Using Undersampling and Oversampling Techniquescited by 5
  18. 2020CNN 101: Interactive Visual Learning for Convolutional Neural Networkscited by 4
  19. 2024MOUNTAINEER: Topology-Driven Visual Analytics for Comparing Local Explanationscited by 4
  20. 2024A Large-Scale Sensitivity Analysis on Latent Embeddings and Dimensionality Reductions for Text Spatializationscited by 4
  21. 2016Learning Interpretability for Visualizations using Adapted Cox Models through a User Experimentcited by 3
  22. 2020Communicative Visualizations as a Learning Problemcited by 3
  23. 2022ENS-t-SNE: Embedding Neighborhoods Simultaneously t-SNEcited by 3
  24. 2019Sea of Genes: Combining Animation and Narrative Strategies to Visualize Metagenomic Data for Museumscited by 2
  25. 2020A Visual Analytics Approach for Hardware System Monitoring with Streaming Functional Data Analysiscited by 2
  26. 2022A Machine-Learning-Aided Visual Analysis Workflow for Investigating Air Pollution Datacited by 1
  27. 2022MetaStackVis: Visually-Assisted Performance Evaluation of Metamodelscited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 7.

8 papers.

  1. 2025Consensus dimension reduction via multi-view learningcited by 1
  2. 2026TopoPrune: Robust Data Pruning via Unified Latent Space Topologycited by 1
  3. 2026LLM-Augmented Semantic Steering of Text Embedding Projection Spacescited by 1
  4. 2026IPBC: An Interactive Projection-Based Framework for Human-in-the-Loop Semi-Supervised Clustering of High-Dimensional Datacited by 0
  5. 2026Class Angular Distortion Index for Dimensionality Reductioncited by 0
  6. 2026MEDAL: Manifold Embedding Distillation via Autoencoder Learningcited by 0
  7. 2026Manifold Bandits: Bayesian Curriculum Learning over the Latent Geometry of Large Language Modelscited by 0
  8. 2026Understanding Differentiable Embeddings Through Differential and Integral Geometrycited by 0