Learning Sleep Stages from Radio Signals: A Conditional Adversarial Architecture

Mingmin Zhao (1)   Shichao Yue (1)   Dina Katabi (1)   Tommi Jaakkola (1)   Matt Bianchi (2)

1Massachusetts Institute of Technology
2Massachusetts General Hospital

Overview:
RF-Sleep learns to predict sleep stages from radio measurements without any attached sensors on subjects. We introduce a new predictive model that combines convolutional and recurrent neural networks to extract sleep-specific subject-invariant features from RF signals and capture the temporal progression of sleep. A key innovation underlying our approach is a modified adversarial training regime that discards extraneous information specific to individuals or measurement conditions, while retaining all information relevant to the predictive task.
Video:

Paper:
Learning Sleep Stages from Radio Signals: A Conditional Adversarial Architecture
Mingmin Zhao, Shichao Yue, Dina Katabi, Tommi Jaakkola, Matt Bianchi
International Conference on Machine Learning (ICML’17)
[PDF]

Talk:

Dataset:
This dataset contains RF measurements during sleep with corresponding sleep stage labels.