This page is available in German language only.
Thank you very much for your interest in writing a master’s thesis in collaboration with DATEXIS.
When deciding on a topic, please refer to focus areas and our current publications.
Selection process. The first step is a short meeting for coordination, for example after the lecture "Enterprise Data Management". If the topic is suitable, we encourage you to prepare a two-page exposé. In this document, you present your idea, the context within our research areas, relevant literature, and your methodology. After submitting the exposé, we jointly decide whether your work fits the focus areas of DATEXIS.
Next steps. If there is a "go", we recommend participating in our regular colloquia, where students exchange ideas with each other. You should plan approximately 3 months (working three to four days per week) for designing the concept, implementation, and evaluation/testing of your system, and about 2 months for writing the thesis.
For structuring and organizing your thesis, we provide an orientation guide.
Featured Theses
Sebastian von Rohrscheidt. Evaluation reasoning capabilities of language model for clinical outcome prediction. Master Thesis BHT 2025 [pdf][slides]
Dennis Fast. Domain Adaptation of Transformer-based Language Models via Low Layer Information Integration. Master Thesis BHT 2023 [pdf][slides]
Jan Frick. Kidney Graft Loss Prediction leverag- ing Multimodal Health Records. Master Thesis BHT 2022 [pdf][slides]
Philipp Haustein. Transformer-based Medication Prediction from Clinical Notes. Bachelorarbeit BHT 2021 [pdf][slides]
Anjali Grover. Multi-task Learning with AdapterFusion. Masterarbeit BHT 2021. [pdf][slides]
Rouven Reuter. Prediction of Permanent Kidney Graft Loss via Deep Ensemble Learning and NLP using a biased Dataset of Electronic Health Records. Masterarbeit BHT 2021. [pdf][slides]
Arndt Alhorn. Transfer Learning for Hate Speech Detection. Masterarbeit BHT 2020. Ausgezeichnet als bester Absolvent des Sommersemester 2020 im Studiengang Data Science (Master) [pdf][slides]
Sebastian Jäger. Compressing BERT - An Evaluation and Combination of Methods. Masterarbeit BHT 2020. [pdf][slides]
Ivana Trajanovska. Assertion Detection in Electronic Health Records. Masterarbeit BHT 2020. [pdf][slides]
Alexei Figueroa. Knowledge enhanced language models. Masterarbeit BHT 2020. [pdf][slides]
Paul Grundmann. Reading Long Healthcare Documents with Neural Transformers. Masterarbeit BHT 2020. [pdf][slides]
Shuaib Yunus. Learning to Rank Places Masterarbeit Beuth-Hochschule 2019.
[pdf][slides]
Thomas Steffek. Neural Facet Detection on Medical Resources. Bachelorarbeit Beuth-Hochschule 2019. [pdf][slides] Ausgezeichnet als bester Absolvent des Wintersemester 2018/2019 im Studiengang Medieninformatik (Bachelor)
Christopher Kümmel. Learning a Sampling Strategy for Named Entity Recognition. Masterarbeit Beuth-Hochschule 2018. [PDF][PDF]
Cordula Guder. Visual Analysis of Structural Documents (mit SAP Innovation Center). Ausgezeichnet als beste Absolventin des Wintersemester 2017/2018 im Studiengang Medieninformatik (Master)
Denis Martin. Konzeptbezogenes Entity Linking auf Dokumentenebene mit Deep Learning. Masterarbeit Beuth-Hochschule 2017. [PDF]
Lukas Abegg. Suchoptimierung mittels maschinellem Lernen (mit SpringerNature). Bachelorarbeit 2016 [PDF][PPT] Ausgezeichnet als beste Bachelorarbeit des Sommersemester 2017 im Studiengang Medieninformatik (Bachelor)
Tom Oberhauser. Eingrenzung der geographischen Region nicht verorteter Tweets anhand von Text und Metadatenanalyse. Bachelorarbeit Beuth-Hochschule 2016. Ausgezeichnet als bester Absolvent des Wintersemester 2016/2017 im Studiengang Medieninformatik (Bachelor) [PPT] [PDF] [Ehrung durch den Dekan]
Fabian Grimme: Learning to Recognize Annotations. Masterarbeit Beuth-Hochschule 2016. [PDF] [PPT]
Claudia Schaefer: Interaktive Sentimentklassifikation für Textdaten. (Interactive Sentiment Classification for Text Data) Master Thesis Beuth-Hochschule, Ausgezeichent als beste Absolventin des Sommersemesters 2015 im Studiengang Medieninformatik (Master) [DEMO1] bzw. [DEMO2]
Alexander Muschalle: Preismodelle auf Datenmarktplätzen (Pricing Models on Data Marketplaces). Bachelorarbeit TU Berlin 2013. Published in "Pricing Approaches for Data Markets". BIRTE Workshop with VLDB 2012: 129-144
Sebastian Arnold: GoOLAP User Interaktion. Bachelorarbeit TU Berlin, 2011. Published in Lecture Notes in Business Information Processing Volume 96, Springer Verlag 2012
Christoph Boden: Classification Algorithms for Web Text
Filtering. Diplomarbeit TU Berlin, 2011. Published in ICDE Workshops 2011.
Martin Bach: Extraktion komplexer Relationen aus englischsprachigen Webseiten (Extraction of Complex Relations from English Websites). Diplomarbeit TU Berlin, 2010.
Selected Graduates
… and their first employers can be found here.
