Embeddings & Vector SearchHervé Jégou et al. (INRIA)
Jégou et al. publish Product Quantization for Nearest Neighbor Search
Published Product Quantization for Nearest Neighbor Search — decomposed the space into a Cartesian product of low-dimensional subspaces to compress vectors into short codes, evaluated at a scale of 2 billion vectors
Why it matters
The origin point of the IVF-PQ family. The compressed indexes of FAISS, ScaNN and most vector DBs descend directly from this paper
Source — Primary sources first; secondary reports are labelled “Press”
Officialinria.hal.sciencehttps://inria.hal.science/inria-00514462Added in Week 35, 2026 · See that week