< Back to previous page

Project

Contextual Anomaly Detection for Complex Industrial Assets

CONSCIOUS studies the detection of abnormal behavior in industrial machines and processes. These applications contain sensors that send out data, and that data is searched for irregularities. What is innovative about this research, is its focus on context. After all, in some contexts, certain deviations can be explained perfectly and thus are not so alarming. CONSCIOUS takes into account disruptive contextual factors and thus wants to offer solutions for more accurate anomaly detection in complex and heterogeneous data. The project teaches AI systems to search for the causes of irregularities in contextually enriched data. In other words, AI shows when an irregularity is contextual and when something is really going on. The results will be validated by real-world industrial use cases in different domains.

Date:1 Jan 2021 →  30 Jun 2023
Keywords:AI, anomaly detection, irregularities in contextually enriched data
Disciplines:Machine learning and decision making