Vector Search in Modern Databases: From Traditional to AI-Powered Information Retrieval

Explore the transformative world of vector search in modern databases through this 25-minute conference talk from OSACon. Delve into how traditional databases and search engines like Lucene, Elasticsearch, Solr, PostgreSQL, MySQL, MongoDB, and Manticore are incorporating vector search capabilities, marking a significant evolution in information retrieval. Learn fundamental concepts of vectors and embeddings, understand various embedding computation strategies including in-database processing and external tools, and master different indexing approaches for optimizing vector search performance. Discover effective hybrid search techniques that combine vector-based and traditional search methods, while gaining insights into measuring both performance and quality metrics. Conclude with a forward-looking perspective on how vector search technologies are shaping the future of data retrieval and AI applications.