GMV presents how to detect industrial anomalies with GMV PitIA at the TDWI Conference in Munich

GMV participated in the TDWI Conference in Munich, one of Europe’s leading events for data, analytics, and artificial intelligence, with Miguel Tejedor and Fernando López presenting the paper “Explainable Unsupervised Anomaly Detection for Telemetry” on how to bring artificial intelligence from laboratories into the day-to-day operations of complex industrial systems.

During the session, these GMV experts demonstrated the capabilities of GMV PitIA, a solution designed to convert large volumes of industrial telemetry into useful information for operational decision-making. The proposal is based on a unified model capable of learning the normal behavior of processes from latent multivariate representations of correlated signals, making it possible to address different use cases within a single technological framework.

Notable capabilities presented included the prediction of process variables using virtual sensors, unsupervised anomaly detection without the need for pre-labeled data, event-based operational monitoring, and the identification of the variables responsible for each detected deviation, facilitating root cause analysis and the interpretation of results by operators. The system also incorporates continuous adaptation mechanisms that keep the models aligned with the evolution of industrial processes.

The presentation also highlighted the importance of combining explainable models, event-level reasoning, and continuous updating to reduce false alarms, improve incident management, and facilitate the integration of artificial intelligence into existing platforms and operational workflows. In addition, GMV PitIA can adapt to both scenarios with large volumes of historical data and environments where such data is very limited—as is that case in space applications involving satellites at the start of their operations—thanks to an approach that reduces reliance on large training datasets. This flexibility makes it easy to deploy across a wide range of industrial and mission environments.

 GMV PitIA reaffirms the company’s commitment to developing reliable, explainable, and operation-oriented artificial intelligence solutions capable of helping organizations transform data into faster, more accurate, and more actionable decisions in highly critical industrial environments.