AGENDA

14:00-14:15: Introduction of the AI program and the session - Sabine Demey (imec)

14:15-14:45: Anomaly detection in time series data - Roberto Medico (UGent - imec)

14:45-14:50: Testimonial by industrial partner - Kristof Coddens (Melexis)

14:50-15:00: Q&A about the talk with the attendees

ABSTRACT talk Roberto Medico (PhD Student in Machine Learning & Data Analysis (UGent - imec).

Anomaly detection for time-series data is a critical and challenging task that is relevant in many domains, including e.g. manufacturing processes, healthcare, financial transactions.

This webinar will guide the audience through the rationale, mechanism and performance of time-series anomaly detection methods based on machine learning. Both shallow and deep learning-based approaches will be highlighted.

This will be illustrated on a use-case in the electrical domain, where the detection of errors and faults is important to evaluate the robustness of electrical devices against noise and interference.

Tickets are free but registration is mandatory. We hope to see you the 17th of November. If you already have some questions, feel free to ask.

Kind regards

The Flanders AI Research Team