What is approximate nearest neighbor (ANN) search and why is it used?

Approximate nearest neighbor search, or ANN, finds vectors close to a query vector without guaranteeing exact top-k results, trading a small amount of recall for massive speedups. Exact search requires comparing the query to every vector in the index, which is linear in dataset size and becomes impractical beyond a few hundred thousand vectors. ANN algorithms like HNSW, IVF, and DiskANN build index structures that achieve logarithmic or sublinear query time while holding 95 to 99 percent recall in production settings.