TransferLab Training: Practical Anomaly Detection
This workshop introduces unsupervised ML techniques for anomaly detection, their strengths and weaknesses and different application areas.
To improve the comprehension of the presented material, we adapted the course to be in a flipped classroom setting. The course is spread over 7 weeks in which you’ll work through the material in your own pace and have the chance to ask questions and discuss with our experts in weekly meetings. Also, feel free to discuss on relevant research papers and ideas.
Identification of anomalies is of high interest in multiple industries for identifying safety and security risks, ensuring production quality, or finding new business opportunities. But anomalies are hard to detect because they are characteristically poorly represented in datasets.
The training covers the following topics:
- An informal notion of anomaly and the types of anomaly detection like unsupervised learning, one class problem, class imbalances, etc.
- Density estimation and robustness for anomaly detection
- Anomaly detection via isolation and reconstruction errors
- Anomaly detection on time series
- Extreme value theory for anomaly detection
All topics are presented from a theoretical perspective and applied to illustrative problems.
Starting with a kick-off event, the training spans a total of 7 weeks in which the participants go over the course material at their own pave and discuss the topics with the experts in weekly meetings.
More information about this training’s content can be found at our TransferLab website.
Disclaimer
The appliedAI Institute for Europe gGmbH is supported by the KI-Stiftung Heilbronn gGmbH.
The appliedAI Institute for Europe gGmbH is a subsidiary of the appliedAI Initiative GmbH.
