Logo des Repositoriums

Auflistung nach Autor:in "Netz, Lukas"

1 - 9 von 9
  • Workshopbeitrag
    Combining 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, Bernhard
    We 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to produce models for well-known modeling languages such as PlantUML, however, they perform worse on lesser-known or unpublished DSLs.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.
  • Zeitschriftenartikel
    Enterprise Information Systems in Academia and Practice - Lessons learned from a MBSE Project.
    (EMISA Forum: Vol. 39, No. 1, 2019) Adam, Kai; Michael, Judith; Netz, Lukas; Rumpe, Bernhard; Varga, Simon
  • Konferenzbeitrag
    Enterprise Information Systems in Academia and Practice: Lessons learned from a MBSE Project
    (40 Years EMISA 2019, 2020) Adam, Kai; Michael, Judith; Netz, Lukas; Rumpe, Bernhard; Varga, Simon
    The development of domain-specific information systems, especially web information systems, takes a certain amount of time, needs intensive testing to ensure a certain quality and lacks the consistency of front- and backend. Using model-based strategies for the creation of information systems helps to overcome these problems by fastening the development process, facilitating testing and ensuring consistency-by-construction. In practice, however, they are still rarely used. In this paper, we show that model-based engineering is beneficial for the creation of an enterprise information system and improves the quality of the resulting product. We present the basic functionalities of our Generator for Enterprise Management (MontiGEM) and discuss identified problems and lessons learned in a project in practice. The generator was developed simultaneously with and for an enterprise management system. Our research shows that the use of generative methods and MBSE improves the adaptability and reusability of parts of the application on the one hand but on the other hand, there are still obstacles that slow down its broad application in practice.
  • Konferenzbeitrag
    From Natural Language to Web Applications: Using Large Language Models for Model-Driven Software Engineering
    (Modellierung 2024, 2024) Netz, Lukas; Michael, Judith; Rumpe, Bernhard
    We evaluate the usage of Large Language Models (LLMs) to transform natural language into models of a predefined domain-specific language within the context of model-driven software engineering. In this work we test systematically the reliability and correctness of the developed tooling, to ensure its usability in an automated model-driven engineering context. Up to now, LLMs such as ChatGPT were not sophisticated enough to yield promising results. The new API-Access and the release of GPT-4, enabled us to develop improved tooling that can be evaluated systematically. This paper introduces an approach that can produce a running web application based on simple informal specifications, that is provided by a domain expert with no prior knowledge of any DSL. We extended our toolchain to include ChatGPT and provided the AI with additional DSL-specific contexts in order to receive models that can be further processed. We performed tests to ensure the semantic and syntactic correctness of the created models. This approach shows the potential of LLMs to successfully bridge the gap between domain experts and developers and discusses its current limitations.
  • Konferenzbeitrag
    Generating Digital Twin Cockpits for Parameter Management in the Engineering of Wind Turbines
    (Modellierung 2022, 2022) Michael, Judith; Nachmann, Imke; Netz, Lukas; Rumpe, Bernhard; Stüber, Sebastian
    The complexity of wind energy systems combined with an increased trend towards mass customization require the collaboration of many experts to achieve high quality products. Currently, a major issue arises from the lack of data integration among the different tools used during the engineering process which may cause system failures eventually. Existing tools largely do not support automatic detection and indication of erroneous or contradictory parameter values between artifacts of different tools. Employing a model-driven and functional engineering approach enables to establish an integrated toolchain for the management and visualization of engineering artifacts that consume and produce the data. Within this paper, we present an automatic approach to derive an engineering digital twin for the cooperative development and management of engineering artifacts from functional models of the system under development. We evaluate our approach on the example of a hydraulic pump within the cooling system of a wind turbine. The prototype can be coupled with an existing engineering tool ecosystem. The approach enables to exchange the data produced by engineering artifacts according to a functional system model which facilitates the cooperation between different stakeholders throughout the development process.
  • Zeitschriftenartikel
    Model-Based Generation of Enterprise Information Systems.
    (EMISA Forum: Vol. 38, No. 1, 2018) Adam, Kai; Netz, Lukas; Varga, Simon; Michael, Judith; Rumpe, Bernhard; Heuser, Patricia; Letmathe, Peter
  • Konferenzbeitrag
    Model-Based Software Engineering at RWTH Aachen University
    (40 Years EMISA 2019, 2020) Adam, Kai; Michael, Judith; Netz, Lukas; Rumpe, Bernhard; Varga, Simon
    In this article, the Software Engineering Research Group of RWTH Aachen University presents its research aim, research topics and some research highlights that have been researched over the last ten years. Furthermore, the relevance of agile, generative and model-based software engineering methods for the special interest group Enterprise Modelling and Information Systems Architectures (SIG-EMISA) is discussed.
  • Konferenz-Abstract
    Retrofitting Generative Aspects in Existing Applications
    (EMISA 2022, 2022) Gerasimov, Arkadii; Michael, Judith; Nachmann, Imke; Netz, Lukas; Rumpe, Bernhard; Varga, Simon
  • Zeitschriftenartikel
    Towards Privacy-Preserving IoT Systems Using Model Driven Engineering.
    (EMISA Forum: Vol. 40, No. 1, 2020) Michael, Judith; Netz, Lukas; Rumpe, Bernhard; Varga, Simon