Latent Stochastic Differential Equations for Change Point Detection

Automated analysis of complex systems based on multiple readouts remains a challenge. Change point detection algorithms are aimed to locating abrupt changes in the time series behaviour process. In this paper, we present novel change algorithm Latent Neural Stochastic Differential Equations (SDE). Our method learns non-linear deep learning transformation process into latent space and estimates SDE that describes its evolution over time. The uses likelihood ratio learned stochastic processes different timestamps find points We demonstrate capabilities performance our synthetic real-world datasets. proposed outperforms state-of-the-art majority experiments.