Disentangled Adversarial Transfer Learning for Physiological Biosignals
This addresses transfer learning challenges in wearable sensor-based physiological monitoring, offering a method to adapt models to broader subject ranges, though it appears incremental as it builds on existing adversarial and disentanglement techniques.
The paper tackles the problem of domain inconsistency in physiological biosignals across users or sessions for stress assessment, proposing an adversarial inference approach that extracts disentangled nuisance-robust representations, with results showing benefits in cross-subject transfer evaluations.
Recent developments in wearable sensors demonstrate promising results for monitoring physiological status in effective and comfortable ways. One major challenge of physiological status assessment is the problem of transfer learning caused by the domain inconsistency of biosignals across users or different recording sessions from the same user. We propose an adversarial inference approach for transfer learning to extract disentangled nuisance-robust representations from physiological biosignal data in stress status level assessment. We exploit the trade-off between task-related features and person-discriminative information by using both an adversary network and a nuisance network to jointly manipulate and disentangle the learned latent representations by the encoder, which are then input to a discriminative classifier. Results on cross-subjects transfer evaluations demonstrate the benefits of the proposed adversarial framework, and thus show its capabilities to adapt to a broader range of subjects. Finally we highlight that our proposed adversarial transfer learning approach is also applicable to other deep feature learning frameworks.