PROJECT / 01
Connect questions to the institute’s knowledge.
A research project to query LARIT’s knowledge in natural language using an ontology and language models.
- ROLE
- AI research and development
- SCOPE
- Ontology · Graph reasoning · Evaluation
- STATUS
- In development · interface planned
- YEAR
- 2026

CONTEXT
Make research easier to explore
An institute’s publications, researchers and projects form a network of connected knowledge. ARIS explores how to query it in natural language while retaining links to the data supporting the answer.
APPROACH
Structure, experiment, compare
The work combines an RDF/OWL ontology with experiments connecting LLMs to knowledge graphs, including GCR, ChatKBQA and ORT. Decisions are documented and answers evaluated on question sets. Constrained reasoning aims to reduce unsupported answers; it does not guarantee error-free output.
PROGRESS
Research in progress
The ontology, experimental implementations and evaluations are the core of the project at this stage. The final user interface remains to be built. The portfolio visual is an illustrative diagram, not an application screenshot.
NEXT PROJECT
E-OPS