Anomaly detection is an essential task for quality management in smart manufacturing. An accurate data-driven detection method usually needs enough data and labels. However, in practice, there commonly exist newly set-up processes in manufacturing, …