PET: An Annotated Dataset for Process Extraction from Natural Language Text Tasks

Process extraction from text is an important task of process discovery, for which various approaches have been developed in recent years. However, differently from other information extraction tasks, there is a lack of gold-standard corpora of business process descriptions carefully annotated with all the entities and relationships of interest. This paper presents the PET dataset, a first corpus of business process descriptions annotated with activities, gateways, actors, and flow information. We present our new resource, including a variety of baselines to benchmark the difficulty and challenges of business process extraction from text. The PET dataset, annotation guidelines, and inception schema are freely available via huggingface.co/datasets/patriziobellan/PET.