Developer Voices Semantic Search: A Deep Dive Into Vector Databases (with Zain Hasan)
10 snips
Oct 18, 2023 
Zain Hasan, a semantic search and augmented LLMs expert, joins the podcast to discuss the challenges of teaching large language models. They explore the concept of vector databases and their role in enhancing chat bots. The episode delves into optimizing search in a fictional service, the size and storage of indexes in vector databases, and the concept of multi-modality in vector search. The discussion also includes implementing semantic search at home using Weaviate and a conversation on Weaviate, an open-source database with managed instances.
AI Snips
Chapters
Transcript
Episode notes
Semantic Search via Vector Distances
- Vector search finds documents with vectors closest to a query vector using distance metrics like Euclidean distance.
- Sorting by smallest distances returns the most semantically similar entries.
Approximate Nearest Neighbors Scaling
- Brute force nearest neighbors search is slow and scales linearly with data size and vector dimensions.
- Vector databases require approximate nearest neighbors algorithms to scale to millions of vectors efficiently.
Trade Accuracy for Speed
- Use approximate nearest neighbors algorithms to trade slight accuracy loss for vast performance gains.
- This enables real-time vector searches over billions of documents.