We are extremely excited to announce the release of metaphactory 5.1, which comes with improved model governance, as well as multiple improvements to the next-gen search.
Improved model governance through provenance information
Aiming to streamline collaboration in semantic knowledge modeling, metaphactory 5.1 enables its users to automatically capture provenance information for ontologies and vocabularies on any creation or edit operation.
Provenance data is directly captured as annotation metadata to the ontology or vocabulary and to the individual element that is being edited (e.g., a class, a relation, an attribute, a term, etc.). This is achieved by utilizing open and well-known metadata standards like Dublin Core. While this feature will immediately enhance traceability and improve the collaboration process, it is only the first step in a series of upcoming features to improve the model governance.
New features in next-gen search
Vocabularies and their respective hierarchical structures have always been supported in metaphactory's search features, for example, through hierarchical facets for refinement of query results. With the latest enhancements in metaphactory 5.1, we deliver improved support for utilizing vocabularies in the query formulation process by allowing end users to access and exploit vocabulary structures when visually defining their query. End users can expand their search results to access more sources that might be relevant, while at the same time keeping full control over the search precision.
Additionally, metaphactory 5.1 enables end users to formulate more expressive queries by allowing them to explore a selection of multiple keyword result entities within next-gen search.
What else is new in metaphactory 5.1?
- Natural Language Interaction - Integration of Large Language Models (LLMs): metaphactory 5.1 delivers a first blueprint integration with LLMs, demonstrating how generative AI models can support end users in using natural language questions to interact with knowledge graphs. The blueprint, implemented as a metaphactory app, integrates initially with the public version of ChatGPT. It allows end users to formulate information needs using natural language queries, which are translated to SPARQL queries and executed over the factual knowledge in the knowledge graph. For the time being, the app is available on request for beta testing only.
- Improved language handling and the ability to enforce the ordering of languages for semantic forms
- Access to the latest Highcharts 11 features in semantic charts, incl. new chart types like (inverted) tree graph
- Support for editing SKOS mapping properties through the vocabulary editor
- SSO Azure On-Behalf token exchange for token delegation to the database (GraphDB, Stardog)
- Improved log management
- Improved default typography and legibility
Have a look at the release notes or