Concept · Building with AI

Similarity search

Also called: nearest neighbor search, nearest neighbour search, k-nearest neighbors, kNN, ANN search, approximate nearest neighbor

Finding the stored items whose vectors are closest to a query's vector, so results match by meaning rather than exact words. Vector databases are built to do it quickly.

For example: A shopper searches 'something warm for winter', and similarity search returns coats and sweaters though neither word was typed.

Connects to 7

• Vector database Concept
• Cosine similarity Concept
• Embedding Concept
• Lexical search Concept
• Reranking Concept
• Recommendation engine Concept
• Hybrid search Concept