pgvector - Stylish Hierarchical Navigable Small World Indexes

Explore the world of vector similarity search in this seminar talk from the CMU Database Group's ML⇄DB Series. Dive into pgvector and its implementation of Hierarchical Navigable Small World (HNSW) indexes, presented by Jonathan Katz from Amazon and PostgreSQL. Learn about the challenges and solutions in developing efficient vector search capabilities for PostgreSQL, including the design decisions, performance optimizations, and future directions for this powerful extension. Gain insights into how pgvector enhances database functionality for machine learning applications and vector-based operations. Suitable for database professionals, researchers, and anyone interested in the intersection of databases and machine learning.

pgvector: Stylish Hierarchical Navigable Small World Indexes (Jonathan Katz)