_Semantic search_, or vector-based retrieval, as apposed to [[Lexical Search]], is a type of search that seeks to understanding the context of the query and find related documents. Once a search application can map documents and queries to dense vectors, retrieval with [[Cosine Similarity]] is fairly straightforward. See [[Information Retrieval (IR)]] Read: https://softwaredoug.com/blog/2023/02/13/why-vector-search ## [Background](https://www.sbert.net/examples/applications/semantic-search/README.html#background "Permalink to this headline") The idea behind semantic search is to embed all entries in your corpus, whether they be sentences, paragraphs, or documents, into a vector space. At search time, the query is embedded into the same vector space and the closest embeddings from your corpus are found. These entries should have a high semantic overlap with the query.  [[Symmetric & Asymmetric Semantic Search]]