Bi-Encoder
Model that embeds queries and passages separately so they can be compared cheaply.
Overview
A bi-encoder encodes the query and each passage independently into vectors, and relevance is their similarity. It solves the problem that a cross-encoder must run once per candidate, which is impossible against a large collection. It exists as the half of retrieval that can be precomputed and indexed, at the cost of losing the interaction between query and passage.
Technology Graph
How Bi-Encoder connects to the rest of the stack.
Applied to
- AI Search application Applications Bi-Encoder used for AI Search High ✓ source
- Embedding concept Techniques Bi-Encoder used for Embedding High ✓ source
Tools
Software that implements or integrates with Bi-Encoder.
Frameworks
Alternatives
-
Cross-Encoder
concept
An alternative to Bi-Encoder · source
Comparisons
Side-by-side breakdowns with the alternatives above.