Measuring Quality and Interpretability of Dimensionality Reduction Visualizations

Research output: Contribution in Book/Catalog/Report/Conference proceedingConference contribution

3 Downloads (Pure)

Abstract

One first step to get insights about a dataset can be its visualization using dimensionality reduction (DR). However, DR processes induce a loss of information that needs to be quantified in order to evaluate the quality of their results. Furthermore, two DR visualizations with a similar loss value can be really different in the eyes of the user. This paper presents DR quality measures developed in the machine learning community, as well as visual quality measures considered in the information visualization community, which can be used to assess interpretability. We propose to combine several measures from these two categories in order to be able to predict and study users' understanding of DR visualizations.
Original languageEnglish
Title of host publicationSafeML ICLR Workshop
Place of PublicationNew Orleans, Louisiana
Publication statusPublished - 2019

Fingerprint

Visualization
Learning systems

Keywords

  • Machine learning
  • Interpretability
  • Dimensionality reduction
  • Quality metrics

Cite this

@inproceedings{e9f3668bc7f8415fa4c92b49f82858fc,
title = "Measuring Quality and Interpretability of Dimensionality Reduction Visualizations",
abstract = "One first step to get insights about a dataset can be its visualization using dimensionality reduction (DR). However, DR processes induce a loss of information that needs to be quantified in order to evaluate the quality of their results. Furthermore, two DR visualizations with a similar loss value can be really different in the eyes of the user. This paper presents DR quality measures developed in the machine learning community, as well as visual quality measures considered in the information visualization community, which can be used to assess interpretability. We propose to combine several measures from these two categories in order to be able to predict and study users' understanding of DR visualizations.",
keywords = "Machine learning, Interpretability, Dimensionality reduction, Quality metrics",
author = "Adrien Bibal and Beno{\^i}t Frenay",
year = "2019",
language = "English",
booktitle = "SafeML ICLR Workshop",

}

Measuring Quality and Interpretability of Dimensionality Reduction Visualizations. / Bibal, Adrien; Frenay, Benoît.

SafeML ICLR Workshop. New Orleans, Louisiana, 2019.

Research output: Contribution in Book/Catalog/Report/Conference proceedingConference contribution

TY - GEN

T1 - Measuring Quality and Interpretability of Dimensionality Reduction Visualizations

AU - Bibal, Adrien

AU - Frenay, Benoît

PY - 2019

Y1 - 2019

N2 - One first step to get insights about a dataset can be its visualization using dimensionality reduction (DR). However, DR processes induce a loss of information that needs to be quantified in order to evaluate the quality of their results. Furthermore, two DR visualizations with a similar loss value can be really different in the eyes of the user. This paper presents DR quality measures developed in the machine learning community, as well as visual quality measures considered in the information visualization community, which can be used to assess interpretability. We propose to combine several measures from these two categories in order to be able to predict and study users' understanding of DR visualizations.

AB - One first step to get insights about a dataset can be its visualization using dimensionality reduction (DR). However, DR processes induce a loss of information that needs to be quantified in order to evaluate the quality of their results. Furthermore, two DR visualizations with a similar loss value can be really different in the eyes of the user. This paper presents DR quality measures developed in the machine learning community, as well as visual quality measures considered in the information visualization community, which can be used to assess interpretability. We propose to combine several measures from these two categories in order to be able to predict and study users' understanding of DR visualizations.

KW - Machine learning

KW - Interpretability

KW - Dimensionality reduction

KW - Quality metrics

M3 - Conference contribution

BT - SafeML ICLR Workshop

CY - New Orleans, Louisiana

ER -