Embedding
Dense numeric vector representing the meaning of a piece of text.
Overview
An embedding maps a piece of text to a fixed-length vector of numbers positioned so that similar meanings land close together. It solves the problem that keyword matching cannot tell that two differently worded sentences mean the same thing. It exists because nearest-neighbour search over vectors is fast and can be indexed, which makes semantic retrieval practical at scale. Embeddings are what make retrieval work at all: the query is embedded, and the closest stored documents are returned as candidates.
Technology Graph
How Embedding connects to the rest of the stack.
Uses and depends on
- Recommendation application Applications Recommendation based on Embedding High ✓ source
Used by
- AI Search application Applications Embedding used by AI Search High ✓ source
- Agent Memory technique AI Agents Embedding used by Agent Memory High ✓ source
- Chroma database Vector Databases Embedding used by Chroma High ✓ source
- Elasticsearch database Search Engines Embedding used by Elasticsearch High ✓ source
- GraphRAG technique Techniques Embedding used by GraphRAG High ✓ source
- HyDE technique Techniques Embedding used by HyDE High ✓ source
- Hybrid Search technique Techniques Embedding used by Hybrid Search High ✓ source
- LanceDB database Vector Databases Embedding used by LanceDB High ✓ source
- Milvus database Vector Databases Embedding used by Milvus High ✓ source
- OpenSearch database Search Engines Embedding used by OpenSearch High ✓ source
- Pinecone database Vector Databases Embedding used by Pinecone High ✓ source
- Qdrant database Vector Databases Embedding used by Qdrant High ✓ source
- RAG technique Techniques Embedding used by RAG High ✓ source
- Redis database Databases Embedding used by Redis High ✓ source
- Typesense database Search Engines Embedding used by Typesense High ✓ source
- Vespa database Search Engines Embedding used by Vespa High ✓ source
- Weaviate database Vector Databases Embedding used by Weaviate High ✓ source
- pgvector library Vector Databases Embedding used by pgvector High ✓ source
Applied to
- BEIR benchmark Benchmarks BEIR used for Embedding High ✓ source
- Bi-Encoder concept Techniques Bi-Encoder used for Embedding High ✓ source
- Chunking technique Techniques Chunking used for Embedding High ✓ source
- Contrastive Learning technique Algorithms Contrastive Learning used for Embedding High ✓ source
- HNSW algorithm Algorithms HNSW used for Embedding High ✓ source
- MS MARCO dataset Datasets MS MARCO used for Embedding High ✓ source
- MTEB benchmark Benchmarks MTEB used for Embedding High ✓ source
- SQLite database Databases SQLite used for Embedding High ✓ source
- Sentence Transformers framework Frameworks Sentence Transformers used for Embedding High ✓ source
- Tokenizer tool Infrastructure Tokenizer used for Embedding High ✓ source