Why it matters
PQ enables billion-scale ANN. Understanding shapes vector DB.
Advertisement
The architecture
Split d-dim vector into m subvectors.
Quantize each with k centroids.
Store bytes-per-subvector.
Query via lookup tables.
Advertisement
How it works end to end
Jegou et al. 2011.
m = 8-32 typical.
k = 256 (1 byte).
Combined with IVF for scale.