The DAI-Labor offers possible theses for Bachelor of Science (BSc) / Master of Science (MSc) for several of its research foci. For further information and advice on the individual topics, please contact the respective supervisor.

Print Loading print view...

When applying models, it is necessary to make individual decisions and the model interpretable and explainable. White-box models such as linear regression are straightforward to interpret and explain but are limited in their performance, especially if the model's data generating process differs. Therefore, empirical, data-based machine learning models are often used because they require little to no expert knowledge of the data when modeling. However, many of these are grey or black-box models, hence very complex and performant models that cannot (or only sparsely) be explained by the model and do not allow for statistical inference. Their results are not comprehensible, and model behavior cannot be explained.

In Explainable AI (XAI), a subfield of Artificial Intelligence, methods are developed to make such models comprehensible and interpret individual decisions. In addition to the development of XAI methods, suitable visualizations of interpretations need to be developed, for example, for complex time series data, e.g., multivariate time series.

The goal of this thesis is to find suitable visualizations of interpretations for different XAI methods for (multivariate) time series forecasts. The aim is to develop visualizations that allow the comparability of interpretations across different forecast models or XAI methods.

This is a bachelor thesis topic.


  • Experience with time series analysis and prediction, good comprehension of mathematical/statistical basics are helpful
  • Experience with visualization of time series is desirable
  • Very good comprehension of Machine Learning, first experiences with XAI methods are advantageous
  • Good programming skills (ideally Python)


  • Guidotti, Riccardo & Monreale, Anna & Turini, Franco & Pedreschi, Dino & Giannotti, Fosca. (2018). A Survey of Methods for Explaining Black Box Models. ACM Computing Surveys. 51. 10.1145/3236009.
  • W. Samek, G. Montavon, S. Lapuschkin, C. J. Anders and K. -R. Müller, "Explaining Deep Neural Networks and Beyond: A Review of Methods and Applications," in Proceedings of the IEEE, vol. 109, no. 3, pp. 247-278, March 2021, doi: 10.1109/JPROC.2021.3060483.
  • Schlegel, Udo & Arnout, Hiba & El-Assady, Mennatallah & Oelke, Daniela & Keim, Daniel. (2019). Towards A Rigorous Evaluation Of XAI Methods On Time Series. 4197-4201. 10.1109/ICCVW.2019.00516.

research area

Machine Learning
supervisor / Contact person
Print Loading print view...

Predictive maintenance refers to the prediction of errors and defects in hardware components before these errors occur. This can be used to obtain replacement parts early and avoid major subsequent failures.

Machine learning algorithms have a great potential for predictive maintenance, as machine learning can be used to discover structures in maintenance data and in this way predict a maintenance data time series. In doing so, imminent defects can be detected earlier by using established physical reasoning to find the defects using the data. At this point, there is potential for improvement by not only limiting a system to known rationales and also performing the defect detection itself using machine learning techniques.

The aim of this work is to investigate how machine learning based algorithms for time series prediction can be used, possibly combined with other methods, to do predictive maintenance.


  • Very good knowledge about analysis, linear algebra and optimization
  • Interest and experience in machine learning
  • good programming skills (ideally Python)


  • Fernandes, M., Canito, A., Corchado, J. M., & Marreiros, G. (2019, June). Fault detection mechanism of a predictive maintenance system based on autoregressive integrated moving average models. In International Symposium on Distributed Computing and Artificial Intelligence (pp. 171-180). Springer, Cham.
  • Kanawaday, A., & Sane, A. (2017, November). Machine learning for predictive maintenance of industrial machines using IoT sensor data. In 2017 8th IEEE International Conference on Software Engineering and Service Science (ICSESS) (pp. 87-90). IEEE.
  • Yang, H., Pan, Z., Tao, Q., & Qiu, J. (2018). Online learning for vector autoregressive moving-average time series prediction. Neurocomputing, 315, 9-17.

research area

Machine Learning
supervisor / Contact person
Lukas Friedemann Radke
Print Loading print view...

In the domain of medical image analysis for pathology (histopathology), there is a huge potential for applying modern machine learned image analysis models on tasks like segmentation of cell nuclei or classification of nuclei by cell type. In scientific studies, these models seem to perform well and therefore, one would like to apply them in medical practice. One of the many requirements for that is a very careful and well executed validation of their generalization performance on benchmark datasets. However, because of the very expensive procedure of obtaining images for a validation set and several factors that create interdependencies between the images that violate the common iid assumption of many machine learning approaches, specific care must be taken when validating such models.

The goal of this thesis is to explore the effects of specific violations of the iid-assumption with respect to validating the generalization performance of machine learned models for histopathological images.


  • Very solid knowledge in statistics and validation of machine learning for image based tasks (Machine Learning 1, Data Science or similar courses)
  • Interest in bridging the gap between theoretically oriented research and solid practical outcomes
  • Good programming skills in Python or languages with strong deep learning frameworks
  • Knowledge about the specific medical domain is helpful, but not mandatory


  • Homeyer, A. et. Al. (2021). Artificial Intelligence in Pathology: From Prototype to Product. Journal of Pathology Informatics (DOI 10.4103/jpi.jpi_84_20)
  • He, Y. Shen, Z. Cui, P. (2021). Towards non-iid image classification: A dataset and baselines. Pattern Recognition Volume 110 ISSN 0031-3203

research area

Machine Learning
supervisor / Contact person
Christian Geißler
Print Loading print view...

Time Series such as sensor signals, speech or stock prices generally vary in length and speed. To cope with such temporal variations, warping distances such as Dynamic Time Warping (DTW) are often used. In recent years, the problem of finding an average of time series under DTW has been faced and many properties such as the existence, uniqueness, complexity, exact solutions and heuristic solutions have been studied. In practice, a DTW average may not 'look like' an averagely shaped curve of the sample time series. This is reasoned in the nature of the DTW distance itself. Therefore, several alternative warping distances have been proposed. However, as of today, they have rarely been used for computing average curves. The goal of this thesis it to explore the behavior of different warping distances for time series averaging. One challenge is to evaluate such averages quantitatively and qualitatively.


  • Good programming skills (Python, Matlab or Java)
  • Solid basics in Analysis and Linear Algebra, ideally experience with gradient descent optimizers
  • Interest in theoretically oriented research


  • Schultz, D., & Jain, B. (2018). Nonsmooth analysis and subgradient methods for averaging in dynamic time warping spaces. Pattern Recognition74, 340-358.
  • Petitjean, F., Ketterlin, A., & Gançarski, P. (2011). A global averaging method for dynamic time warping, with applications to clustering. Pattern recognition44(3), 678-693.
  • Zhao, J., & Itti, L. (2018). shapedtw: Shape dynamic time warping. Pattern Recognition74, 171-184.
  • Keogh, E. J., & Pazzani, M. J. (2001, April). Derivative dynamic time warping. In Proceedings of the 2001 SIAM international conference on data mining (pp. 1-11). Society for Industrial and Applied Mathematics.
  • Cuturi, M., & Blondel, M. (2017, July). Soft-dtw: a differentiable loss function for time-series. In International Conference on Machine Learning (pp. 894-903). PMLR.
  • Marteau, P. F. (2008). Time warp edit distance with stiffness adjustment for time series matching. IEEE transactions on pattern analysis and machine intelligence31(2), 306-318.
  • And own literature research
supervisor / Contact person
M.Sc. David Schultz
Print Loading print view...

Geodesically convex optimization has recently attracted attention from the machine learning community due to the realization that some important problems that appear to be non-convex at first glance, are geodesically convex, if we introduce a suitable differential structure and a metric. In this thesis, the student is expected to conduct an empirical study on the performance of geodesic convex optimization methods for applications in machine learning, compared to the state-of-the-art non-convex optimization techniques.


  • English language
  • successfully completed courses Machine Learning I and II
  • Programming skills in Python
  • Knowledge in Differential Geometry and Convex Optimisation
supervisor / Contact person
Print Loading print view...

The Electrical Grid plays an important role in our everyday life. As to bring the new technologies in place and facilitate the energy transition, Simulations have been conducted to assess the feasibility of new developments. Today we do not only talk about the stability, reliability and efficiency of the grid but also its resilience or the ability to Anticipate, React, Adapt, and Recover. Therefore, to increase the resilience of today’s power grid, simulation and software assessments are needed to couple the multidomain inter-dependencies of the Cyber-Physical Human Systems (CPHS). Extreme weather events have been occurring more frequently around the world. These events demonstrated how vulnerable the distribution grid is.

Well-placed and coordinated enhancements, such as system hardening and redundancy, orchestrated microgrids, and advanced fault detection, can reduce the number of outages that could occur due to High Impact Low Probability (HILP) events. We offer a Bachelor / Master Thesis in the field of Power Grid Simulation and integration of Distributed Energy Resources. The thesis will encapsulated the different stages from Model Development to Vulnerability Assessment, and Power Flow Optimization related to Distributed Energy Resources in Power Networks.


  • Studies in the fields of technical informatics, control engineering, automation, electrical engineering, energy engineering, or similar
  • Good knowledge and practical experience in python, co-simulation paradigms, ArcGIS, Simulink or Power Factory
  • Teamwork and Goal Oriented
  • Interest in Renewable Energy Topics and Automation Technology
  • Task Oriented and Independent Learning

research area

Energy Data Analytics
supervisor / Contact person
M.Sc. Izgh Hadachi