Universal Physiological Representation Learning with Soft-Disentangled Rateless Autoencoders

Mo Han, Ozan Özdenizci, Toshiaki Koike-Akino, Ye Wang, Deniz Erdogmus

Research output: Contribution to journalArticle


Human computer interaction (HCI) involves a multidisciplinary fusion of technologies, through which the control of external devices could be achieved by monitoring physiological status of users. However, physiological biosignals often vary across users and recording sessions due to unstable physical/mental conditions and task-irrelevant activities. To deal with this challenge, we propose a method of adversarial feature encoding with the concept of a Rateless Autoencoder (RAE), in order to exploit disentangled, nuisance-robust, and universal representations. We achieve a good trade-off between user-specific and task-relevant features by making use of the stochastic disentanglement of the latent representations by adopting additional adversarial networks. The proposed model is applicable to a wider range of unknown users and tasks as well as different classifiers. Results on cross-subject transfer evaluations show the advantages of the proposed framework, with up to an 11.6% improvement in the average subject-transfer classification accuracy.
Original languageEnglish
Number of pages10
JournalIEEE Journal of Biomedical and Health Informatics
Early online date3 Mar 2021
Publication statusE-pub ahead of print - 3 Mar 2021


  • adversarial learning
  • autoencoders
  • Bioinformatics
  • Biomedical monitoring
  • Decoding
  • deep learning
  • disentangled representation
  • Feature extraction
  • physiological biosignals
  • Physiology
  • soft disentanglement
  • stochastic bottleneck
  • Stochastic processes
  • Task analysis

ASJC Scopus subject areas

  • Health Information Management
  • Electrical and Electronic Engineering
  • Biotechnology
  • Computer Science Applications

Fields of Expertise

  • Human- & Biotechnology
  • Information, Communication & Computing

Fingerprint Dive into the research topics of 'Universal Physiological Representation Learning with Soft-Disentangled Rateless Autoencoders'. Together they form a unique fingerprint.

Cite this