Researchers have developed a novel AI pipeline designed to automatically convert natural language descriptions of business processes into Business Process Model and Notation (BPMN) diagrams. This multi-layered system integrates transformer-based architectures, large language models, and ontology population to address challenges like linguistic ambiguity and the detection of non-adjacent control-flow relationships. The pipeline achieved a 90.7% accuracy in relation extraction and demonstrated strong structural alignment with reference BPMN models in various evaluation scenarios, including a real-world healthcare process. AI
IMPACT This AI pipeline could streamline the creation of business process documentation, improving efficiency and clarity in business operations.
RANK_REASON This is a research paper detailing a novel AI pipeline for a specific task. [lever_c_demoted from research: ic=1 ai=1.0]
- Alexandru Miron Rancea
- arXiv
- Business Process Model and Notation
- DagsHub
- Hugging Face
- large-language models
- Roberta
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →