The resulting corpus can be traversed by policy, by entity, by lineage, and by semantic similarity rather than by similarity alone so that retrieval uses graph structure in addition to text resemblance.

The result is a corpus that can be traversed by policy, by entity, by lineage, and by semantic similarity

— Beyond Retrieval - Pyrana Context Engine Whitepaper.pdf

rather than by

— Beyond Retrieval - Pyrana Context Engine Whitepaper.pdf

similarity alone.

— Beyond Retrieval - Pyrana Context Engine Whitepaper.pdf