The Linux Foundation Projects
Skip to main content

Vector Search

In information retrieval, vector search represents data (documents, images, etc.) as vectors in a multi-dimensional space, capturing their semantic features. A search query is transformed into a vector, and the search process involves identifying vectors closest to the query, using measures like cosine similarity or Euclidean distance. This method excels in delivering contextually relevant results in large datasets, forming the backbone of many contemporary search engines and recommendation systems.