From Documents to Database: Failure Modes for Industrial Assets
This addresses the need for efficient structured content creation in enterprise asset management systems, though it appears incremental as it applies existing foundation models to a specific domain.
The paper tackles the problem of generating Failure Mode and Effects Analyses (FMEA) for industrial equipment by using foundation models and technical documents, resulting in reduced time for creating this knowledge-intensive content compared to manual approaches.
We propose an interactive system using foundation models and user-provided technical documents to generate Failure Mode and Effects Analyses (FMEA) for industrial equipment. Our system aggregates unstructured content across documents to generate an FMEA and stores it in a relational database. Leveraging this tool, the time required for creation of this knowledge-intensive content is reduced, outperforming traditional manual approaches. This demonstration showcases the potential of foundation models to facilitate the creation of specialized structured content for enterprise asset management systems.