Artwork

177: Vector Databases

Programming Throwdown

224 subscribers

published

iconShare
 
Manage episode 448487414 series 2417399
Content provided by Patrick Wheeler and Jason Gauci, Patrick Wheeler, and Jason Gauci. All podcast content including episodes, graphics, and podcast descriptions are uploaded and provided directly by Patrick Wheeler and Jason Gauci, Patrick Wheeler, and Jason Gauci or their podcast platform partner. If you believe someone is using your copyrighted work without your permission, you can follow the process outlined here https://staging.podcastplayer.com/legal.

Intro topic: Buying a Car

News/Links:

Book of the Show

Patreon Plug https://www.patreon.com/programmingthrowdown?ty=h

Tool of the Show

Topic: Vector Databases (~54 min)

  • How computers represent data traditionally
    • ASCII values
    • RGB values
  • How traditional compression works
    • Huffman encoding (tree structure)
    • Lossy example: Fourier Transform & store coefficients
  • How embeddings are computed
    • Pairwise (contrastive) methods
    • Forward models (self-supervised)
  • Similarity metrics
  • Approximate Nearest Neighbors (ANN)
  • Sub-Linear ANN
    • Clustering
    • Space Partitioning (e.g. K-D Trees)
  • What a vector database does
    • Perform nearest-neighbors with many different similarity metrics
    • Store the vectors and the data structures to support sub-linear ANN
    • Handle updates, deletes, rebalancing/reclustering, backups/restores
  • Examples
    • pgvector: a vector-database plugin for postgres
    • Weaviate, Pinecone
    • Milvus

★ Support this podcast on Patreon ★
  continue reading

184 episodes