The current prototype (RDF & Rule Workbench) uses an existing RDF 1.2 model (with annotations) and enriches it with rules (in Datalog or SWRL format) by merging the base model with the inferences. The result can be visualized as a graph and validated using SPARQL and SHACL. Specific product details and data from simulated supply chain processes illustrate the relationship and application of the solution in the context of the Digital Product Passport. The current version features a configurable AI search that allows users to ask questions in natural language. A language model translates these into SPARQL queries, which then display the results. For each query, the workbench displays the model’s runtime, the SPARQL query’s runtime, and the number of tokens for input, output, and total. Locally executed models (e.g., Ollama) can also be used for processing product data that is not intended to leave the company.
View of the RDF 1.2 Rule Workbench prototype
Would you like to try out and discuss this semantic AI search (based on an RDF model and, optionally, using Datalog or SWRL rules) with your own product data? Please feel free to contact us.