Master Thesis: Generative Modelling Based Anomaly Detection

[Bas van Stein]

[Posted in name of Lucas Correia]

Our Team is located in the powertrain development department. We develop evaluation systems and calibration methods for powertrain test benches. Every day immense amounts of data are recorded, therefore this data must be of sufficient quality if it is to be used for later analysis, in other words, it must be free of anomalies.

Your tasks in detail:

-Create a generative model (VAE, GAN, ...) of the system being tested

-Find way to use the model in anomaly detection algorithm

-Test the model in offline environment with multivariate time-series data

-Deploy model on test bench computer (or at embedded device, like a Jetson Nano, if needed) and ensure real-time capability

Any questions can be directed to lucas....@daimler.com

 

The posting can be found here: https://jobs.mercedes-benz.com/Stellenanzeige/DE_328328/master-thesis-in-deep-learning-based-anomaly-detection-starting-may-2022.html