Autor*innen mit den meisten Dokumenten
Auflistung nach:
Neueste Veröffentlichungen
- WorkshopbeitragLimitations of ChatGPT in Conceptual Modeling: Insights from Experiments in Metamodeling(Modellierung 2024 Satellite Events, 2024) Muff, Fabian; Fill, Hans-Georg
- WorkshopbeitragPreface for Workshop Research Data Management in Modelling in Computer Science (RDiMOD)(Modellierung 2024 Satellite Events, 2024) Koschmider, Agnes; Terboven, Christian; Goedicke, MichaelPreface of the RDiMOD'24 Workshop
- WorkshopbeitragIntegrating Declarative and Imperative Process Modeling Paradigms in the Age of Generative AI(Modellierung 2024 Satellite Events, 2024) Kampik, Timotheus; Berg, Gregor; Eickhoff, DavidThis brief paper summarizes a talk introducing and discussing the notion of process atoms, small facts or queries, each describing an organizationally relevant property or constraint of a business process that cannot be further split without losing its business meaning. An example of a process atom is: “only if an order with a purchase amount greater than 10,000€ is requested, management approval has to take place afterwards” (more abstractly: “only if A then eventually B”). As process atoms are executable as queries on data and allow for dynamic contextualization across process and organizational scopes, they complement and augment traditional process models, such as BPMN diagrams, particularly in the age of data-driven process analysis and generative AI-created process content.
- WorkshopbeitragTowards theoretical foundations for large process models(Modellierung 2024 Satellite Events, 2024) Fettke, Peter
- WorkshopbeitragLeveraging LLMs in Semantic Mapping for Knowledge Graph-based Automated Enterprise Model Generation(Modellierung 2024 Satellite Events, 2024) Reitemeyer, Benedikt; Fill, Hans-GeorgAutomated enterprise model generation applies artificial intelligence and other machine- processable approaches to improve decision making and adoption in complex and changing en- vironments. The emergence of Large Language Models (LLMs) opens a new playing field for machine-processability in enterprise modeling, especially when it comes to processing natural lan- guage contextual knowledge. In this extended abstract, we show the use of LLMs in semantic mapping tasks for real-world and modeling language concepts based on an ArchiMate and National Information Exchange Model (NIEM) example. The results indicate that LLMs are useful in automated enterprise modeling tasks.
- WorkshopbeitragCombining Retrieval-Augmented Generation and Few-Shot Learning for Model Synthesis of Uncommon DSLs(Modellierung 2024 Satellite Events, 2024) Baumann, Nils; Diaz, Juan Sebastian; Michael, Judith; Netz, Lukas; Nqiri, Haron; Reimer, Jan; Rumpe, BernhardWe introduce a method that empowers large language models (LLMs) to generate models for domain-specific languages (DSLs) for which the LLM has little to no training data on. Common LLMs such as GPT-4, Llama 2, or Bard are trained on publicly available data and thus have the capability to produce models for well-known modeling languages such as PlantUML, however, they perform worse on lesser-known or unpublished DSLs. Previous work focused on the usage of few-shot learning (FSL) to synthesize models but did not address or evaluate the potential of retrieval-augmented generation (RAG) to provide fitting examples for the FSL-based modeling approach. In this work, we propose a toolchain and test each building block individually: We use the MontiCore Sequence Diagram Language, which GPT-4 has minimal training data on, to assess the extent to which FSL enhances the likelihood of synthesizing an accurate model. Additionally, we evaluate how effectively RAG can identify suitable models for user requests and determine whether GPT-4 can distinguish between requests for a specific model and those for general information. We show that RAG and FSL can be used to enable simple model synthesis for uncommon DSLs, as long as there is a fitting knowledge base that can be accessed to provide the needed examples for the FSL approach.
- KonferenzbeitragAutomatisierte Verarbeitung natürlichsprachlich repräsentierter Sachverhalte zur Identifizierung von Kandidaten für Bezeichner in Datenmodellen(Modellierung 2024 Satellite Events, 2024) Christ, Sven; Strecker; StefanFür das Bestimmen von Kandidaten für Bezeichner von Modellelementen (Entitätstypen, Beziehungstypen, Attributen) aus natürlichsprachlich repräsentierten Sachverhaltsbeschreibungen werden für die Datenmodellierung mit der Modellierungssprache „Entity-Relationship Model“ (ERM) Heuristiken vorgeschlagen, die an Morphologie und Grammatik der natürlichen Sprache orientiert sind. Bereits seit den 1990er Jahren werden diese Heuristiken in Verbindung mit Ansätzen des „Natural Language Processing“ (NLP) eingesetzt, um für das Erstellen von Datenmodellen eine (teil-) automatisierte Modellierungsunterstützung zu realisieren. In diesem Beitrag kontrastieren wir die für das Modellierungswerkzeug TOOL implementierte NLP-basierte Modellierungsunterstützung mit drei Transformer-basierten künstlichen neuronalen Netzen, „Large Language Model“ (LLM), hinsichtlich fünf unterschiedlich komplexen Aufgaben des Identifizierens von Kandidaten für Bezeichner von Modellelementen in einer Variante des ERM. Die vorliegenden, noch vorläufigen Ergebnisse deuten an, dass die verwendeten LLM dem kontrastierten regelbasierten NLP-Ansatz deutlich überlegen sind.
- KonferenzbeitragVisualizing Model and Data Differences with Inline Diff Editors in an Enterprise Low-Code Platform(Modellierung 2024 Satellite Events, 2024) Butting, Arvid; Greifenberg, Timo; Hölldobler, Katrin; Kehrer, TimoComputing the difference between models is a foundation for solving various challenges in model-driven engineering. Models of low-code platforms are often of non-textual nature to be suited best for citizen developers as their creators. This poses special challenges for visualizing model differences. In this talk, we extend an approach for visualizing model differences via diff editors derived from the original model editors of the low-code platform by a novel kind of diff representation called inline diff editors. The approach improves on the previous approach through a concise form of presentation and by avoiding the generation of new user interface and data models for the editors.
- KonferenzbeitragMessage from the Modellierung’24 Industry-Forum Chairs(Modellierung 2024 Satellite Events, 2024) Loos, Peter; Proper, Henderik A.Message from the Modellierung’24 Industry-Forum Chairs
- KonferenzbeitragPractical Experience with Petriflow: Enriched Process Models Serving as Implementation(Modellierung 2024 Satellite Events, 2024) Juhás, Gabriel; Mladoniczky, Milan; Petrovič, ĽubošIn this paper we discuss experiences gained by using low-code language Petriflow, based on extended Petri nets enriched by data variables and forms. We illustrate on several real-life use cases how the Petriflow models of business processes can directly be used as implementation when deployed in Petriflow interpreter.
- «
- 1 (current)
- 2
- 3
- »