search

ZJULearning / nsg

Navigating Spreading-out Graph For Approximate Nearest Neighbor Search
MIT License
630 stars 149 forks source link

The distance calculation of inner product is wrong. #37

Open CaucherWang opened 1 year ago

CaucherWang commented 1 year ago

The inner product of two vectors represents the similarity. So we should return the largest ones as kNN. It's opposite to L2 distance.

fc731097343 commented 12 months ago

Currently we only support Euclidean distance. Nearest neighbor search mainly refers to Euclidean space. Maximum inner product problem is another active literature in info retrieval. We will consider scale our work to it in future