The paper explores the integration of Large Language Models (LLMs) into process modeling to automate the generation and refinement of process models from textual descriptions. It proposes a framework leveraging LLMs for model generation, error handling, and feedback incorporation. Preliminary results demonstrate the framework's ability to streamline process modeling tasks, highlighting the transformative potential of generative AI in BPM.
Traditional process modeling methods are time-consuming and require expertise in complex languages like BPMN or Petri nets. The paper introduces a novel framework that leverages LLMs to automate the generation of process models from textual descriptions. This approach aims to make process modeling more accessible to users without expertise in modeling languages.
The framework involves innovative prompting strategies for effective LLM utilization, secure model generation protocols, error-handling mechanisms, and user feedback integration for model refinement. By implementing a concrete system extending this framework, robust quality guarantees on generated models are provided.
Several related works explore different approaches for extracting process information from text using NLP techniques or combining NLP with computational linguistics techniques. Commercial vendors are also integrating AI into process modeling systems.
In eine andere Sprache
aus dem Quellinhalt
arxiv.org
Wichtige Erkenntnisse aus
by Humam Kouran... um arxiv.org 03-13-2024
https://arxiv.org/pdf/2403.07541.pdfTiefere Fragen