Overview

Vector search is a way to find related objects that have similar characteristics using multiple machine learning models that detect semantic relationships between objects in indexes. One way to find similar vectors is to use k-nearest neighbors (kNN) algorithm, which returns the k nearest vectors, by comparing every data vector in the database to the query vector. However, in the real world, the size of data items increases, and the number of dimensions that most models use in similarity search goes up to hundreds or thousands of dimensions as well, thus the brute force of kNN search is computationally very expensive. Instead of using kNN, we usually focus on Approximate Nearest Neighbor (ANN) algorithm, which trade off a bit of accuracy for a huge improvement in speed.