Vector Searchconcept

Retrieval that ranks items by vector similarity (typically approximate nearest neighbor) rather than keyword match.

Embeddings make vector search the natural retrieval primitive for semantic queries; it underpins RAG and clustering.

Where it appears

Descend

Historical evolution

ANN indexes (HNSW, IVF) made billion-scale vector search practical.

Implementation details

SKCE's explorer does NOT do live vector search; the curriculum's search is a pre-built inverted index (compile-time, not runtime).

Sources

View in graph →