AnomalyKiTS: Anomaly Detection Toolkit for Time Series

This demo paper presents a design and implementation of system AnomalyKiTS for detecting anomalies from time series data the purpose offering broad range algorithms to end user, with special focus on unsupervised/semi-supervised learning. Given an input series, provides four categories model building capabilities followed by enrichment module that helps label anomaly. also supports wide execution engines meet diverse need anomaly workloads such as Serveless CPU intensive work, GPU deep-learning training, etc.