Multi-Disciplinary Process Research
Processes are, in a simplified way, a representation of intent. As such, they are a great way to capture the needs of various organizations all while offering opportunities to leverage different types of innovations that target the application domain. In XRCE we do a lot of research in machine learning, computer vision, natural language processing and analytics so it is natural to leverage it to boost our process intelligence methods and tools.
We are working on a number of promising research directions that combine process modeling and management with other such fields. Below is a sample of multi-disciplinary projects we are pursuing:
- Querying and managing process collections using semantic modeling. This involves managing various technical representations of process artefacts (such as BPMN and domain-specific process elements) in ways that ensure consistency rules are not broken, for instance when generated designs are manually updated. In addition, this allows for complex querying of process designs and process collections using domain language constructs. Check out this related paper.
- Computer vision embedding in process designs. This relates to matching computer vision capabilities (such as detection or tracking) in process specifications at the granularity of process tasks. Check out this paper for the description of a first approach centred on integration with BPMN.
- Natural language processing and machine learning for extracting process structure from unstructured textual data.