Finished Theses

More on the formal side

In the context of probabilistic inference

• Tensor network decompositions for HMMs (BA)
• Parallelising Probabilistic Inference in Bayesian Networks (BA)

• Approximating Probabilistic Inference by Treewidth Minimization in Graphical Models (MA)

In the context of decision making

• Investigating Approaches to Model Reconciliation in PGMs for Explainable Decision Support (BA)

• Lifting Multi-agent A* - Implementing a New Optimisation Problem in Isomorphic DecPOMDPs: How many agents do we need? (MA)

• Partitioning and Lifting in Multi-Agent Reinforcement Learning (MA)

More application-oriented

In the context of privacy

• An Implementation and Evaluation of Max Cover Clustering Algorithms in the Differentially Private Framework (MA)

In the context of text modelling

• Topic Modelling of Anthropological Journals Using Latent Dirichlet Allocation (BA)

In the context of databases

• In cooperation with the Institute of Medical Informatics: Steiner Goes Big Data: Validating Database Entries in Large Databases (MA)

Second examination

• Sichere Lösungen der Datenhaltung in Webanwendungen umgesetzt in einem Werkzeug zur Verwaltung von Multiple-Choice-Fragen (MA), Erstgutachter: Ludger Becker

• The Impact of Noisy Observations on Deep Deterministic Policy Gradient (BA), Erstgutachter: Malte Schilling

• Automatic Evaluation Metrics for Natural Language Generation (BA), Erstgutachter: Christian Grimme, Wirtschaftsinformatik

• Practicality of Differentially Private Clustering (BA), first examiner: Esfandiar Mohammadi, University of Lübeck

• Integrating Categorial Data into Star Coordinates (MA), first examiner: Lars Linsen
• Efficiency of Route Planning Algorithms Using Distance Tables (BA), first examiner: Jan Vahrenhold
• Utilising Shapley Values for an explainable neural netwerk in biomedical applications (MA), first examiner: Xiaoyi Jiang
• Development of an Application for Automated Data Quality Analysis - Is a Dataset a Valid Subset of a Database Instance? (MA), first examiner: Ludger Becker
• Supporting the operational verhicle claims settlement by deep learning based detection of car parts and damages (MA), first examiner: Xiaoyi Jiang
• Principal Component Analysis for Explainability of Genetic Test Generation in Satellite Image Processing (MA), first examiner: Paula Herber