Biosignal Generation and Latent Variable Analysis with Recurrent Generative Adversarial Networks

Research output: Contribution to journalArticle

Abstract

The effectiveness of biosignal generation and data augmentation with biosignal generative models based on generative adversarial networks (GANs), which are a type of deep learning technique, was demonstrated in our previous paper. GAN-based generative models only learn the projection between a random distribution as input data and the distribution of training data. Therefore, the relationship between input and generated data is unclear, and the characteristics of the data generated from this model cannot be controlled. This study proposes a method for generating time-series data based on GANs and explores their ability to generate biosignals with certain classes and characteristics. Moreover, in the proposed method, latent variables are analyzed using canonical correlation analysis (CCA) to represent the relationship between input and generated data as canonical loadings. Using these loadings, we can control the characteristics of the data generated by the proposed method. The influence of class labels on generated data is analyzed by feeding the data interpolated between two class labels into the generator of the proposed GANs. The CCA of the latent variables is shown to be an effective method of controlling the generated data characteristics. We are able to model the distribution of the time-series data without requiring domain-dependent knowledge using the proposed method. Furthermore, it is possible to control the characteristics of these data by analyzing the model trained using the proposed method. To the best of our knowledge, this work is the first to generate biosignals using GANs while controlling the characteristics of the generated data.

Original languageEnglish
Article number8794813
Pages (from-to)144292-144302
Number of pages11
JournalIEEE Access
Volume7
DOIs
Publication statusPublished - Jan 1 2019

Fingerprint

Labels
Time series
Deep learning

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Materials Science(all)
  • Engineering(all)

Cite this

Biosignal Generation and Latent Variable Analysis with Recurrent Generative Adversarial Networks. / Harada, Shota; Hayashi, Hideaki; Uchida, Seiichi.

In: IEEE Access, Vol. 7, 8794813, 01.01.2019, p. 144292-144302.

Research output: Contribution to journalArticle

@article{e2a6fc0d533243bb949a6ee57039987b,
title = "Biosignal Generation and Latent Variable Analysis with Recurrent Generative Adversarial Networks",
abstract = "The effectiveness of biosignal generation and data augmentation with biosignal generative models based on generative adversarial networks (GANs), which are a type of deep learning technique, was demonstrated in our previous paper. GAN-based generative models only learn the projection between a random distribution as input data and the distribution of training data. Therefore, the relationship between input and generated data is unclear, and the characteristics of the data generated from this model cannot be controlled. This study proposes a method for generating time-series data based on GANs and explores their ability to generate biosignals with certain classes and characteristics. Moreover, in the proposed method, latent variables are analyzed using canonical correlation analysis (CCA) to represent the relationship between input and generated data as canonical loadings. Using these loadings, we can control the characteristics of the data generated by the proposed method. The influence of class labels on generated data is analyzed by feeding the data interpolated between two class labels into the generator of the proposed GANs. The CCA of the latent variables is shown to be an effective method of controlling the generated data characteristics. We are able to model the distribution of the time-series data without requiring domain-dependent knowledge using the proposed method. Furthermore, it is possible to control the characteristics of these data by analyzing the model trained using the proposed method. To the best of our knowledge, this work is the first to generate biosignals using GANs while controlling the characteristics of the generated data.",
author = "Shota Harada and Hideaki Hayashi and Seiichi Uchida",
year = "2019",
month = "1",
day = "1",
doi = "10.1109/ACCESS.2019.2934928",
language = "English",
volume = "7",
pages = "144292--144302",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Biosignal Generation and Latent Variable Analysis with Recurrent Generative Adversarial Networks

AU - Harada, Shota

AU - Hayashi, Hideaki

AU - Uchida, Seiichi

PY - 2019/1/1

Y1 - 2019/1/1

N2 - The effectiveness of biosignal generation and data augmentation with biosignal generative models based on generative adversarial networks (GANs), which are a type of deep learning technique, was demonstrated in our previous paper. GAN-based generative models only learn the projection between a random distribution as input data and the distribution of training data. Therefore, the relationship between input and generated data is unclear, and the characteristics of the data generated from this model cannot be controlled. This study proposes a method for generating time-series data based on GANs and explores their ability to generate biosignals with certain classes and characteristics. Moreover, in the proposed method, latent variables are analyzed using canonical correlation analysis (CCA) to represent the relationship between input and generated data as canonical loadings. Using these loadings, we can control the characteristics of the data generated by the proposed method. The influence of class labels on generated data is analyzed by feeding the data interpolated between two class labels into the generator of the proposed GANs. The CCA of the latent variables is shown to be an effective method of controlling the generated data characteristics. We are able to model the distribution of the time-series data without requiring domain-dependent knowledge using the proposed method. Furthermore, it is possible to control the characteristics of these data by analyzing the model trained using the proposed method. To the best of our knowledge, this work is the first to generate biosignals using GANs while controlling the characteristics of the generated data.

AB - The effectiveness of biosignal generation and data augmentation with biosignal generative models based on generative adversarial networks (GANs), which are a type of deep learning technique, was demonstrated in our previous paper. GAN-based generative models only learn the projection between a random distribution as input data and the distribution of training data. Therefore, the relationship between input and generated data is unclear, and the characteristics of the data generated from this model cannot be controlled. This study proposes a method for generating time-series data based on GANs and explores their ability to generate biosignals with certain classes and characteristics. Moreover, in the proposed method, latent variables are analyzed using canonical correlation analysis (CCA) to represent the relationship between input and generated data as canonical loadings. Using these loadings, we can control the characteristics of the data generated by the proposed method. The influence of class labels on generated data is analyzed by feeding the data interpolated between two class labels into the generator of the proposed GANs. The CCA of the latent variables is shown to be an effective method of controlling the generated data characteristics. We are able to model the distribution of the time-series data without requiring domain-dependent knowledge using the proposed method. Furthermore, it is possible to control the characteristics of these data by analyzing the model trained using the proposed method. To the best of our knowledge, this work is the first to generate biosignals using GANs while controlling the characteristics of the generated data.

UR - http://www.scopus.com/inward/record.url?scp=85073807712&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85073807712&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2019.2934928

DO - 10.1109/ACCESS.2019.2934928

M3 - Article

AN - SCOPUS:85073807712

VL - 7

SP - 144292

EP - 144302

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

M1 - 8794813

ER -