Teun Kromwijk wins Dow Best OML Master Thesis Award 2022
In his thesis, Teun develops and evaluates a method to detect anomalies in unlabeled multivariate time series.
On the 7th of November, 2022, Teun Kromwijk received the Dow Best OML Master Thesis Award 2022. The jury unanimously decided to grant the 10th edition of this award to Teun, which is accompanied by a check of 1,000 Euros. This yearly award is handed over by a representative of Dow at the Operations Management & Logistics Diploma Ceremony in the Fall, of this year by Michael de Graaf (Senior R&D Leader’).
In the thesis, titled “A Self-Supervised Learning Approach for Anomaly Detection of Industrial Systems”, Teun develops and evaluates a method to detect anomalies in unlabeled multivariate time series. The thesis brings a convincing scientific contribution. State-of-the-art techniques are tested on real-world datasets together with a novel representation technique, to assess the benefits and limitations of the latter. Furthermore, a rigorous and systematic analysis is carried out embracing all steps needed to implement an anomaly detection framework, with in-depth discussions on the obtained results. Such a discussion is aimed not only at providing a quantitative assessment of the tested methods but also at providing some insights on potential causes for the observed differences in performance, with a focus on the reliability of the obtained results.
The developed Self-Supervised Multivariate Anomaly Detection (SSMAD) method for monitoring Industrial Internet of Things (IIoT) sensors is proven to be generic in nature and to detect anomalies quickly. This proposed SSMAD method could help generate insights and take proactive actions to mitigate risks or severity levels in a wider setting, including the situation of Dow. In addition, SSMAD could be utilized to continuously monitor a supply chain and detect anomalies that are essential to maintain resilience and provide increased agility to improve business outcomes.
About the DOW-awards
The jury consisted of Aram de Ruiter, Bao Lin from Dow, Laura Genga, and Alp Akcay from Eindhoven University of Technology. Marco Slikker served as non-voting jury chair. The jury judged the theses on their academic contribution and industrial impact. The students that graduated in the past academic year and received at least a grade of 9 for their projects and theses have been nominated for the award. 9 students satisfied this criterium. Also after another nonregular year with corona hiccups, impacting not only teaching but also graduation projects, the jury concluded that all these 9 theses were of outstanding quality. All of the theses well reflect the objectives of the Operations Management & Logistics program, i.e., to use formal models to analyze, improve, and redesign operational processes.