DimReader

Axis lines that explain non-linear projections

Research output: Contribution to journalArticle

Abstract

Non-linear dimensionality reduction (NDR) methods such as LLE and t-SNE are popular with visualization researchers and experienced data analysts, but present serious problems of interpretation. In this paper, we present DimReader, a technique that recovers readable axes from such techniques. DimReader is based on analyzing infinitesimal perturbations of the dataset with respect to variables of interest. The perturbations define exactly how we want to change each point in the original dataset and we measure the effect that these changes have on the projection. The recovered axes are in direct analogy with the axis lines (grid lines) of traditional scatterplots. We also present methods for discovering perturbations on the input data that change the projection the most. The calculation of the perturbations is efficient and easily integrated into programs written in modern programming languages. We present results of DimReader on a variety of NDR methods and datasets both synthetic and real-life, and show how it can be used to compare different NDR methods. Finally, we discuss limitations of our proposal and situations where further research is needed.

Original languageEnglish (US)
JournalIEEE Transactions on Visualization and Computer Graphics
DOIs
StateAccepted/In press - Aug 19 2018

Fingerprint

Computer programming languages
Visualization

Keywords

  • auto-differentiation
  • Data visualization
  • Dimensionality reduction
  • Encoding
  • Non-linear dimensionality reduction
  • Perturbation methods
  • Sensitivity analysis
  • Tools
  • Visualization

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Computer Graphics and Computer-Aided Design

Cite this

@article{3fb1700427a14c66b1cd4ba504b32a8d,
title = "DimReader: Axis lines that explain non-linear projections",
abstract = "Non-linear dimensionality reduction (NDR) methods such as LLE and t-SNE are popular with visualization researchers and experienced data analysts, but present serious problems of interpretation. In this paper, we present DimReader, a technique that recovers readable axes from such techniques. DimReader is based on analyzing infinitesimal perturbations of the dataset with respect to variables of interest. The perturbations define exactly how we want to change each point in the original dataset and we measure the effect that these changes have on the projection. The recovered axes are in direct analogy with the axis lines (grid lines) of traditional scatterplots. We also present methods for discovering perturbations on the input data that change the projection the most. The calculation of the perturbations is efficient and easily integrated into programs written in modern programming languages. We present results of DimReader on a variety of NDR methods and datasets both synthetic and real-life, and show how it can be used to compare different NDR methods. Finally, we discuss limitations of our proposal and situations where further research is needed.",
keywords = "auto-differentiation, Data visualization, Dimensionality reduction, Encoding, Non-linear dimensionality reduction, Perturbation methods, Sensitivity analysis, Tools, Visualization",
author = "Rebecca Faust and Glickenstein, {David A} and Scheidegger, {Carlos Eduardo}",
year = "2018",
month = "8",
day = "19",
doi = "10.1109/TVCG.2018.2865194",
language = "English (US)",
journal = "IEEE Transactions on Visualization and Computer Graphics",
issn = "1077-2626",
publisher = "IEEE Computer Society",

}

TY - JOUR

T1 - DimReader

T2 - Axis lines that explain non-linear projections

AU - Faust, Rebecca

AU - Glickenstein, David A

AU - Scheidegger, Carlos Eduardo

PY - 2018/8/19

Y1 - 2018/8/19

N2 - Non-linear dimensionality reduction (NDR) methods such as LLE and t-SNE are popular with visualization researchers and experienced data analysts, but present serious problems of interpretation. In this paper, we present DimReader, a technique that recovers readable axes from such techniques. DimReader is based on analyzing infinitesimal perturbations of the dataset with respect to variables of interest. The perturbations define exactly how we want to change each point in the original dataset and we measure the effect that these changes have on the projection. The recovered axes are in direct analogy with the axis lines (grid lines) of traditional scatterplots. We also present methods for discovering perturbations on the input data that change the projection the most. The calculation of the perturbations is efficient and easily integrated into programs written in modern programming languages. We present results of DimReader on a variety of NDR methods and datasets both synthetic and real-life, and show how it can be used to compare different NDR methods. Finally, we discuss limitations of our proposal and situations where further research is needed.

AB - Non-linear dimensionality reduction (NDR) methods such as LLE and t-SNE are popular with visualization researchers and experienced data analysts, but present serious problems of interpretation. In this paper, we present DimReader, a technique that recovers readable axes from such techniques. DimReader is based on analyzing infinitesimal perturbations of the dataset with respect to variables of interest. The perturbations define exactly how we want to change each point in the original dataset and we measure the effect that these changes have on the projection. The recovered axes are in direct analogy with the axis lines (grid lines) of traditional scatterplots. We also present methods for discovering perturbations on the input data that change the projection the most. The calculation of the perturbations is efficient and easily integrated into programs written in modern programming languages. We present results of DimReader on a variety of NDR methods and datasets both synthetic and real-life, and show how it can be used to compare different NDR methods. Finally, we discuss limitations of our proposal and situations where further research is needed.

KW - auto-differentiation

KW - Data visualization

KW - Dimensionality reduction

KW - Encoding

KW - Non-linear dimensionality reduction

KW - Perturbation methods

KW - Sensitivity analysis

KW - Tools

KW - Visualization

UR - http://www.scopus.com/inward/record.url?scp=85052654810&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85052654810&partnerID=8YFLogxK

U2 - 10.1109/TVCG.2018.2865194

DO - 10.1109/TVCG.2018.2865194

M3 - Article

JO - IEEE Transactions on Visualization and Computer Graphics

JF - IEEE Transactions on Visualization and Computer Graphics

SN - 1077-2626

ER -