Sebis Day 2026
Sebis Day 2026 will take place on July 1st 2026 in the FMI building of the Garching campus.
A round of presentations on our latest research will be followed with a discussion and exchange at the sebis Stammtisch.
The industry partners and friends of our chair are invited to join the event to get an overview of the latest research advancements and practical projects taking place at our chair.
Collected presentations from the event will be published at the bottom of this page.
Details
| Date | Wednesday, 1 July 2026 |
| Time | 17:00–19:25 (Workshop) 19:25–20:30 (Stammtisch) |
| Place | Fakultät für Mathematik und Informatik (FMI) TUM School of CIT |
| Room | MI Hörsaal 3 |
| Address | Boltzmannstraße 3, 85748 Garching bei München |
Program of Sebis Day 2026
| Time | Presentation | Speaker |
|---|---|---|
| 17:00 | Welcome
| Prof. Dr Florian Matthes |
| 17:05 | Blockchain-based Prediction Markets We present our research on blockchain-based prediction markets. We study inefficiencies and manipulation risks and explore how blockchain infrastructure can enable more efficient trading and more accurate forecasting. |
Jonas Gebele |
| 17:15 | Demokratie-X We present the initial results of a German research initiative exploring the feasibility of a decentralized news platform that promotes privacy, challenges media power concentration, and upholds democratic values for public welfare. |
Jonas Gebele |
| 17:25 | Enterprise Architecture Management We present our research on Artificial Intelligence in Enterprise Architecture Management. We identify the initial results of AI-readiness of corporations in terms of EAM. We are examining how EAM should be used and what needs to be done in the age of AI regarding communication between departments and business processes within AI-driven companies. |
Marvin Chimeka |
| 17:30 | AI-Based Thumbnail Selection We present the results of our research, conducted in cooperation with RTL Netherlands, on the inclusion of Dutch textual context in the selection of a thumbnail. Selecting a frame that adequately conveys what a video is about in a news/entertainment context is a time-consuming process we aim to better automate with this approach. |
Katharina Sommer |
| 17:40 | Synthetic Data for Domain Adaptation We present a research idea on adapting Large Language Models to specialized domains with sparse data availability, such as the medical domain, using grounded synthetic data. We focus on generating and validating domain-specific data to improve model reliability and performance. | Fabian Karl |
| 17:50 | AI-Assisted Contract Analysis We present results of a research project with SYLVENSTEIN Rechtsanwälte to create a system for AI-assisted legal analysis and correction of German employment contracts using Large Language Models. |
Oliver Wardas |
18:00
| Legal AI Use Case Radar We unveil the 2026 version of the Legal AI Use Case Radar and its companion report, which feature the latest insights on Legal AI trends, challenges, and future outlooks. We focus on the findings of our interview studies, where we talked with over 40 legal professionals in the German-speaking region and beyond. | Oliver Wardas |
18:10
| Pause 20 min Kaffee & Kaltgetränke, Kekse | |
| 18:30 | CausalMosaic Individual scientific publications typically reveal only fragments of complex causal mechanisms. Our CausalMosaic project integrates these scattered findings to construct robust, holistic causal models, enabling targeted interventions for decision-making and providing insights into the 'why' behind the data. |
Jonas Gottal |
| 18:40 | Reducing Manual Effort in Systematic Literature Reviews Producing high-quality systematic reviews is a major challenge in evidence-based medicine. This process involves the screening of hundreds of scientific publications. We investigate how this procedure can be simplified using advanced NLP methods. | Joshua Oehms |
18:50
| Early Detection of Dementia through Language In the AssistD project with medical experts of the TUM university clinic and two industry partners we aim to detect early warning signs of dementia through language by using linguistic and phonetic features. | Alexandre Mercier |
19:00
| From Lab to Practice We present the current status of the TUM Aidvice project founded by Google.org to develop trustworthy AI assistance for patients and medical professionals. | Prof. Dr. Florian Matthes |
19:05
| Improving Patient Care with NLP We demonstrate the current state of our research regarding retrieving, structuring, and answering questions using unstructured patient data. We demonstrate the use of local LLMs in processing German data and German patient data as well as how we can use this data for tasks like summarization and question answering. | Katharina Sommer Tristan Till |
19:15
| We present our work on grounded question answering over electronic health records, exploring how far it can be done wih locally deployed models. | Tejaswi Choppa Ibrahim Yurt |
19:25
| Stammtisch We provide food and beverages and give you the opportunity for networking. |
All participants |
| 20:30 | End of Event |
Arrival by Public Transport
- From Hauptbahnhof: take U4/U5 to Odeonsplatz, then U6 in the direction of Garching-Forschungszentrum
- From the airport: S-Bahn S8 (in the direction of Hauptbahnhof) to the third stop, Ismaning (takes about 13 minutes).
There, take regional bus 230 in the direction of Garching-Forschungszentrum, get off at Boltzmannstraße. - From the S-Bahn area: take any S-Bahn to Marienplatz, then change to U6 in the direction of Garching-Forschungszentrum.
Arrival by Car
- Both from the south (e.g. from Munich) and from the north (e.g. from the airport) preferably take the freeway (A9) to the exit Garching Nord (Exit 70).
- After the exit (either right or left) in the direction of the research institutes, cross the intersection with the B11 and turn into Ludwig-Prandtl-Straße.
Parking spaces are available here on the left.
