Autoencoders, a type of artificial neural network, have emerged as a powerful method for anomaly detection in time series data.
Kwantis R&D is applying these models to proactively identify data anomalies in surface logging sensors data. In fact, data anomalies in such sensors can lead to inaccurate calculations and potentially impact decision-making processes.
The application of anomaly detection methods can give an early alert to raise awareness of a data problem and to address it before critical KPIs computation is performed.
https://kwantis.com/about/research/