Vespa
Search platform combining vector and structured ranking in one query.
Visão geral
Vespa serves vector search, keyword search and structured filtering together, with ranking expressed as a program evaluated over the candidates rather than a fixed score. It solves the problem that real ranking is rarely a single similarity: results are filtered by attributes and scored by a combination of signals. It exists because expressing that as one query avoids over-fetching from a vector index and re-ranking in the application.
Grafo de tecnologias
Como Vespa se liga ao resto do stack.
Usa e depende de
- Embedding concept Techniques Vespa uses Embedding High ✓ source
Aplicado a
- AI Search application Applications Vespa used for AI Search High ✓ source