For the operation of large machines in companies or other critical systems in society, it is usually necessary to record and monitor specic machine or system health indicators over time. In the past, the recorded time series were often evaluated manually or by simple heuristics (such as threshold values) to detect abnormal behavior. With the more recent advances in the fields of ML (machine learning) and AI (articial intelligence), ML-based anomaly detection algorithms are becoming increasingly popular for many tasks such as health monitoring or predictive maintenance.
In our research group we develop new unsupervised anomaly detection approaches. Methods and algorithms that we use are (among others) TCN (Temporal Coherence Networks), LSTM (Long Short-Term Memory) and wavelets.
Application areas: predictive maintenance, health services, ECG.
Project members
|
Markus Thill, TH Köln |
 |
Prof. Dr. Wolfgang Konen, TH Köln |
||
Honors
Publications
2020 |
Time Series Encodings with Temporal Convolutional Networks Inproceedings Vasile, Bogdan Filipic Massimiliano (Hrsg.): 9th International Conference on Bioinspired Optimisation Methods and Their Applications (BIOMA), 2020. |
2019 |
Anomaly Detection in Electrocardiogram Readings with Stacked LSTM Networks Inproceedings á, Petra Barancíkov; Holena, Martin; others, (Hrsg.): Proc. 19th Conference Information Technologies - Applications and Theory (ITAT 2019), 2019, (Best Paper Award). |
2018 |
Online Adaptable Time Series Anomaly Detection with Discrete Wavelet Transforms and Multivariate Gaussian Distributions Forschungsbericht Research Center CIOP (Computational Intelligence, Optimization and Data Mining) TH Köln - University of Applied Science, 2018, (submitted to Archives of Data Sciences, Series A (ECDA'2018), preprint available at http://www.gm.fh-koeln.de/ciopwebpub/Thill18a.d/AoDS2018.pdf). |
2017 |
Anomaly Detection in Time Series with Discrete Wavelet Transforms and Maximum Likelihood Estimation Inproceedings Hoffmann, Frank; Hüllermeier, Eyke (Hrsg.): Proceedings 27. Workshop Computational Intelligence, S. 67-71, Universitätsverlag Karlsruhe, 2017. |
Time Series Anomaly Detection with Discrete Wavelet Transforms and Maximum Likelihood Estimation Inproceedings Valenzuela, Olga; Rojas, Ignacio; others, (Hrsg.): International Work-Conference on Time Series (ITISE2017), 2017. |
Online anomaly detection on the Webscope S5 dataset: A comparative study Inproceedings IEEE Conference on Evolving and Adaptive Intelligent Systems (EAIS 2017), S. 1, Springer 2017. |