School of Mathematics/computer Science/physics

Integral part of the Joint Lab for Artificial Intelligence & Data Science of the Leibniz Institute for Agricultural Engineering and Bioeconomy e. V. and Osnabrück University is a Research Training Group. The associated partners are Agrotech Valley Forum, German Research Center for Artificial Intelligence (DFKI), and Osnabrück University of Applied Sciences. The core objective of the Joint Lab is to develop Artificial Intelligence & Data Science expertise, in particular for agricultural technology systems.

For the Research Training Group, the Joint Lab for Artificial Intelligence & Data Science is looking for a

**Research Assistant (M/F/d)**
**(Salary level E 13 TV-L, 100 %)**:
The position is for a period of four years, starting as soon as possible.

**Your tasks**:

- Conducting scientific research on the topic_ Explainable AI methods for multimodal data in the digital bioeconomy_ with the aim of obtaining a doctorate degree
- Development of novel explainable artificial intelligence methods
- Application of AI and explainable AI methods to diverse bioeconomy use cases
- Analysis of multimodal data, e.g., images, sensor data, genetic data
- Preparation of project reports and scientific publications
- Presentation of project results at conferences and workshops

**Required qualifications**:

- Above-average academic degree (Master's or equivalent) in cognitive science, computer science, mathematics, or related fields of study
- Experience with explainable AI and/or explainable/interpretable machine learning are of advantage, as well as respective frameworks in machine learning/deep learning, in particular: Explainable AI/ML, Graph Neural Networks, Object Detection
- Willingness to conduct real-world experimental deployments
- Programming skills (e.g. in Python) and first experience with ML and corresponding libraries (PyTorch, Tensorflow, NumPy, sklearn, etc.).
- Very good English language skills (written and spoken), German language skills are a plus
- Flexibility, creativity and strong communication skills
- High sense of responsibility, reliability, personal commitment and goal-oriented and independent work as well as scientific ambitions

**Additional Qualifications**:

- Experience with versioning tools, such as Git, and unix-based systems, such as Linux

**We offer**:

- The opportunity to publish your papers in conference and journal publications
- The possibility to obtain a doctorate degree
- A highly motivated and international team as part of the Research Training Group
- Interdisciplinary doctoral supervision ensured by teams from Osnabrück and Potsdam
- Flexible work hours and excellent equipment

A description of the research topic is appended to this posting.

The Joint Lab connects the two locations Osnabrück and Potsdam, a willingness to travel is therefore required. The PhD students are supervised by a team of professors and scientists from Osnabrück and Potsdam.

Reference is made to the possibility of part-time employment.

As a family-friendly university, the University of Osnabrück is committed to the compatibility of work/study and family.

The Osnabrück University particularly wants to promote the professional equality of women and men. Therefore, it strives to increase the proportion of the gender that is underrepresented in the respective field.

Severely handicapped applicants or persons of equal status will be given preferential consideration in the event of equal suitability.

Title: Explainable AI Methods for Multimodal Data in the Bioeconomy

Team: Martin Atzmüller (UOS) and Marina Höhne (ATB)

Abstract:
Desired skills of the applicant:
Must: Excellent Knowledge in ML, especially Deep Learning, and in at least one of the following topics: Explainable AI, Graph Neural Networks, Object Detection

References:

- Leonid Schwenke and Martin Atzmueller (2021) Constructing global coherence representations: Identifying interpretability and coherences of transformer attention in time series data. In Proc. IEEE International Conference on Data Science and Advanced Analytics, DSAA 2021, Porto, Portugal, October 6-9, 2021, pages 1-12. IEEE
- Hedström, A., Weber, L., Krakowczyk, D., Bareeva, D., Motzkus, F., Samek, W.,... & Höhne, M. M. C. (2023). Quantus: An Explainable AI Toolkit for Responsible Evaluation of Neural Network Explanations and Beyond. Journal of Machine Learning Research, 24(34), 1-11.
- Gautam, S., Boubekki, A., Hansen, S., Salahuddin, S. A., Jenssen, R., Höhne, M. M., & Kampffmeyer, M. (2022) ProtoVAE: A Trustworthy Self-Explainable Prototypical Variational Model. In Advances in Neural Information Processing Systems.