Student Research Assistant (m/f/d) – Hybrid Machine Learning Models in Additive Manufacturing

Field of study.: Automation technology, business administration, business administration t.o., electrical engineering, computer science, cybernetics, logistics, aerospace engineering, mechanical engineering, mathematics, mechatronics, physics, control engineering, software design, software engineering, technical computer science, technology management, surveying technology, industrial engineering or comparable.

The advertised student research assistant position is part of the DigiAutoFab project, in which both the IPA and the IFF are consortium partners. The IPA develops solutions for the production sector of industrial companies. The Additive Manufacturing (AM) department has extensive expertise in powder bed-based AM and plant technology, as well as analytics, which can be used to validate data generated through simulation. The IFF conducts research in the areas of artificial intelligence (AI) and machine learning (ML) for cognitive production systems. The focus is on hybrid learning methods, i.e., the integration of domain knowledge with data-driven methods. The advertised position is located at the consortium partner IFF.

Within the DigiAutoFab project, the goal of the IFF is to predict the relationships between various parameters in a chemical post-processing procedure of additively manufactured components in order to identify the most suitable input parameters. Since data from the process are only available to a limited extent, hybrid ML approaches that combine data with domain knowledge are to be investigated and further developed. The domain knowledge is available in the form of differential equations (DEs), or initially to be modeled. Therefore, Physics-Informed Neural Networks (PINNs) are to be used in particular for efficient problem-solving.

The position offers an exciting opportunity to delve into the topics of hybrid ML algorithms and post-AM processes. It involves the statistical analysis of experimentally collected data, collaborating in the implementation of various ML algorithms for predicting quality data of the post-AM process. Initially, data-driven and domain knowledge-based approaches (especially PINNs) will be investigated separately and then combined in a hybrid ML approach. Additionally, the position includes setting up and maintaining the necessary hardware.