bencoleman

4 ms·

1. 1.▲
   by bencoleman 5y ago

   Author here - thanks for the feedback everybody! Agree that the weighted version and the skeleton version would be cool to talk about. Perhaps in future :)
2. 2.▲
   by bencoleman 5y ago

   yes, this is true. In 2D/3D, sorting into regular space partitions of exponentially decreasing size (e.g. KD-trees) is far more effective - roughly log(N) vs N^rho (rho < 1). In high dimensions, the simple partitioning strategy runs
3. 3.▲
   by bencoleman 5y ago

   This thread contains many excellent points, and it's true that LSH is no longer SOTA for ANN problems. Generally, LSH indices are much faster to construct but slower to query than other ANN methods (graph-based, cluster-based, etc). In