V-awake: a visual analytics approach for correcting sleep predictions from deep learning models

Research output: Contribution to journalConference articleAcademicpeer-review

Abstract

The usage of deep learning models for tagging input data has increased over the past years because of their accuracy and high performance. A successful application is to score sleep stages. In this scenario, models are trained to predict the sleep stages of individuals. Although their predictive accuracy is high, there are still misclassifications that prevent doctors from properly diagnosing sleep-related disorders. This paper presents a system that allows users to explore the output of deep learning models in a real-life scenario to spot and analyze faulty predictions. These can be corrected by users to generate a sequence of sleep stages to be examined by doctors. Our approach addresses a real-life scenario with absence of ground truth. It differs from others in that our goal is not to improve the model itself, but to correct the predictions it provides. We demonstrate that our approach is effective in identifying faulty predictions and helping users to fix them in the proposed use case.

Keywords

  • Deep Learning
  • Visualization
  • Visual analytics
  • sleep staging

Cite this

@article{749b05336297446da782678fcfd97a7c,
title = "V-awake: a visual analytics approach for correcting sleep predictions from deep learning models",
abstract = "The usage of deep learning models for tagging input data has increased over the past years because of their accuracy and high performance. A successful application is to score sleep stages. In this scenario, models are trained to predict the sleep stages of individuals. Although their predictive accuracy is high, there are still misclassifications that prevent doctors from properly diagnosing sleep-related disorders. This paper presents a system that allows users to explore the output of deep learning models in a real-life scenario to spot and analyze faulty predictions. These can be corrected by users to generate a sequence of sleep stages to be examined by doctors. Our approach addresses a real-life scenario with absence of ground truth. It differs from others in that our goal is not to improve the model itself, but to correct the predictions it provides. We demonstrate that our approach is effective in identifying faulty predictions and helping users to fix them in the proposed use case.",
keywords = "Deep Learning, Visualization, Visual analytics, sleep staging",
author = "{Garcia Caballero}, Humberto and Michel Westenberg and Binyam Gebre and {van Wijk}, Jack",
year = "2019",
month = "3",
day = "21",
doi = "10.1111/cgf.13667",
language = "English",
volume = "38",
pages = "1--12",
journal = "Computer Graphics Forum",
issn = "0167-7055",
publisher = "Wiley-Blackwell",
number = "3",

}

V-awake : a visual analytics approach for correcting sleep predictions from deep learning models. / Garcia Caballero, Humberto; Westenberg, Michel; Gebre, Binyam; van Wijk, Jack.

In: Computer Graphics Forum, Vol. 38, No. 3, 21.03.2019, p. 1-12.

Research output: Contribution to journalConference articleAcademicpeer-review

TY - JOUR

T1 - V-awake

T2 - Computer Graphics Forum

AU - Garcia Caballero,Humberto

AU - Westenberg,Michel

AU - Gebre,Binyam

AU - van Wijk,Jack

PY - 2019/3/21

Y1 - 2019/3/21

N2 - The usage of deep learning models for tagging input data has increased over the past years because of their accuracy and high performance. A successful application is to score sleep stages. In this scenario, models are trained to predict the sleep stages of individuals. Although their predictive accuracy is high, there are still misclassifications that prevent doctors from properly diagnosing sleep-related disorders. This paper presents a system that allows users to explore the output of deep learning models in a real-life scenario to spot and analyze faulty predictions. These can be corrected by users to generate a sequence of sleep stages to be examined by doctors. Our approach addresses a real-life scenario with absence of ground truth. It differs from others in that our goal is not to improve the model itself, but to correct the predictions it provides. We demonstrate that our approach is effective in identifying faulty predictions and helping users to fix them in the proposed use case.

AB - The usage of deep learning models for tagging input data has increased over the past years because of their accuracy and high performance. A successful application is to score sleep stages. In this scenario, models are trained to predict the sleep stages of individuals. Although their predictive accuracy is high, there are still misclassifications that prevent doctors from properly diagnosing sleep-related disorders. This paper presents a system that allows users to explore the output of deep learning models in a real-life scenario to spot and analyze faulty predictions. These can be corrected by users to generate a sequence of sleep stages to be examined by doctors. Our approach addresses a real-life scenario with absence of ground truth. It differs from others in that our goal is not to improve the model itself, but to correct the predictions it provides. We demonstrate that our approach is effective in identifying faulty predictions and helping users to fix them in the proposed use case.

KW - Deep Learning

KW - Visualization

KW - Visual analytics

KW - sleep staging

UR - http://www.scopus.com/inward/record.url?scp=85070077998&partnerID=8YFLogxK

U2 - 10.1111/cgf.13667

DO - 10.1111/cgf.13667

M3 - Conference article

VL - 38

SP - 1

EP - 12

JO - Computer Graphics Forum

JF - Computer Graphics Forum

SN - 0167-7055

IS - 3

ER -