Milvus - Accelerating Approximate Nearest Neighbor Search for Large Scale Datasets

Overview

Explore techniques for accelerating Approximate Nearest Neighbor Search (ANNS) on large-scale datasets in this 37-minute conference talk by Jun Gu from Zilliz. Learn about Milvus, an open-source vector database, and its role in unlocking the potential of unstructured data for AI applications. Discover the evolution of data management in Milvus, including improvements in IVF, IVF Flat, IVF SQ, and IVF PQ. Gain insights into real-world applications such as intelligent writing assistants, image search for company trademarks, and pharmaceutical molecule analysis. Understand the importance of vector databases in flow-based AI applications and unstructured data services for AI.

Syllabus

Intro
Speaker bio
Zilliz: Who we are
Unlock the treasure of unstructured data
The flow-based Al applications
The unstructured data service (UDS) for Al
Vectors are different
Milvus: The big picture
The ANN benchmark
Boost ANN search performance
Data management: before 0.11.0, IVF
Data management: New in 0.11.0, IVF Flat
Data management: New in 0.11.0, IVF SQ, IVF PQ
Our journey
Current progress
Intelligent writing assistant
Image search for company trademark
Pharmaceutical molecule analysis
Welcome to join the Milvus forum