Login / Signup

Wearable Data From Subjects Playing Super Mario, Taking University Exams, or Performing Physical Exercise Help Detect Acute Mood Disorder Episodes via Self-Supervised Learning: Prospective, Exploratory, Observational Study.

Filippo CorponiBryan M LiGerard AnmellaClàudia Valenzuela-PascualAriadna MasIsabella PacchiarottiMarc ValentíIria GrandeAntonio BenabarreMarina GarrigaEduard VietaAllan H YoungStephen M LawrieHeather C WhalleyDiego Hidalgo-MazzeiAntonio Vergari
Published in: JMIR mHealth and uHealth (2024)
We showed that SSL, a paradigm where a model is pretrained on unlabeled data with no need for human annotations before deployment on the supervised target task of interest, helps overcome the annotation bottleneck; the choice of the pretraining surrogate task and the size of unlabeled data for pretraining are key determinants of SSL success. We introduced E4mer, which can be used for SSL, and shared the E4SelfLearning collection, along with its preprocessing pipeline, which can foster and expedite future research into SSL for personal sensing.
Keyphrases
  • electronic health record
  • machine learning
  • big data
  • endothelial cells
  • liver failure
  • intensive care unit
  • physical activity
  • hepatitis b virus
  • sleep quality
  • rna seq
  • deep learning
  • induced pluripotent stem cells