In this article, we propose a software-driven drift detection and calibration framework based on probabilistic observation in latent space using variational autoencoders (VAEs). The proposed method ...