Unsupervised Contrastive Learning for Robust RF Device Fingerprinting Under Time-Domain Shift

The author introduces a novel solution using contrastive learning to address domain shift challenges in RF device fingerprinting, resulting in significant accuracy improvements over baseline models.