Workshop Modeling in the Age of Large Language Models (LLM4Modeling)
Organizers: Peter Fettke (DFKI and Saarland University ), Hans-Georg Fill (University of Fribourg), Julius Köpke (University of Klagenfurt)
Date: 12 March 2024
Location: HPI Potsdam, Germany
Registration: (tbd)
Deadline for Submissions: 22 January 2024 (extended)
Notification: 31 January 2024
Camera-ready versions: 15 Feb 2024
Submissions via EasyChair:https://easychair.org/conferences/?conf=llm4modeling
Large language models (LLM) have received enormous attention in practice and science since ChatGPT at the latest. It is obvious that the use of LLM has the potential to quickly develop some rudimentary aspects of a domain model. However, it is unclear in which precision and quality this might be possible. The question is how modeling deals with LLM in the future. Which influence will LLM have on modeling? How will the tasks of modeling change? Will modeling lose its importance? Or, the other way around, will modeling increase its importance in the future?
The workshop aims at illuminating potentials and challenges of LLM in modeling. The following topics are addressed by the workshop (list is not exhaustive):
- What are the potentials of LLM for modeling?
- How are modeling results evaluated with LLM?
- How are LLM embedded in well-known tools?
- How can LLM be used in teaching for modeling?
- What new modeling tasks are emerging? How do modeling tasks change?
Program (tentative)
| Tuesday March 12, 2024 - Room Wien, HPI Potsdam | Presentation Title |
| 14:00 - 14:30 | Modeling meets Large Language Models (Martin Forell and Selina Schueler) |
| 14:30 - 14:50 | Towards theoretical foundations for large process models (Peter Fettke) |
| 14:50 - 15:10 | Integrating Declarative and Imperative Process Modeling Paradigms in the Age of Generative AI (Timotheus Kampik, Gregor Berg and David Eickhoff) |
| 15:10 - 15:30 | Leveraging LLMs in Semantic Mapping for Knowledge Graph-based Automated Enterprise Model Generation (Benedikt Reitemeyer and Hans-Georg Fill) |
| 15:30 - 16:00 | Coffee Break |
| 16:00 - 16:30 | Combining Retrieval-Augmented Generation and Few-Shot Learning for Model Synthesis of Uncommon DSLs (Lukas Netz, Judith Michael, Bernhard Rumpe, Juan Sebastian Diaz, Haron Nqiri, Nils Baumann and Jan Reimer) |
| 16:30 - 16:50 | Limitations of ChatGPT in Conceptual Modeling: Insights from Experiments in Metamodeling (Fabian Muff and Hans-Georg Fill) |
| 16:50 - 17:00 | Closing Discussion and Publication Initiatives |
Submissions
There are three types of submissions possible: Full papers, short papers and talk abstracts. Full paper submissions can be up to 15 pages long, including the list of references; short papers up to 10 pages and talk abstracts only consist of a title and an extended abstract of a talk (max. 2 pages). All submissions have to conform to the LNI LaTeX template. All contributions will be reviewed by at least 2 members of the program committee; accepted submissions will be published in a joint conference proceedings in the Digital Library of the German Informatics society (GI e. V.) https://dl.gi.de/.
Program Committee:
Balbir Barn, Middlesex University, UK
Bernhard Rumpe, RWTH Aachen, Germany
Chiara Di Francescomarino, University of Trento, Italy
Felix Härer, University of Fribourg, Switzerland
Kurt Sandkuhl, University of Rostock, Germany
Ulrich Frank, University of Duisburg-Essen, Germany
Patrick Delfmann, University of Koblenz, Germany
Pnina Soffer, University of Haifa, Israel
Reinhard Schütte, University of Duisburg-Essen, Germany
Stefanie Rinderle-Ma, Technical University of Munich, Germany