Giters

rapidsai / raft

RAFT contains fundamental widely-used algorithms and primitives for machine learning and information retrieval. The algorithms are CUDA-accelerated and form building blocks for more easily writing high performance applications.

Home Page:https://docs.rapids.ai/api/raft/stable/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

[FEA] Provide an argument to knn algorithms to compute k-farthest neighbors

cjnolet opened this issue · comments

commented

As the title suggests, for some APIs like brute-force this should be as simple as flipping the direction of the k-select. For the fused algos, it'll be more complicated but not by much.

Computing the k-farthest neighbors with a max spanning tree is also going to be needed in order to compare the complete linkage hierarchical clustering, which is an algo we aim to support at some point.

commented

Hi, @cjnolet This issue is about the construction of k-farthest neighbor graph on a close-set dataset, or perform a k-farthest neighbors search given some queries ?

commented

@JieFengWang this issue is about allowing kselection to be flipped so that return the largest distances instead of the smallest.

commented

@JieFengWang this issue is about allowing kselection to be flipped so that return the largest distances instead of the smallest.

Hi, @cjnolet what is k-select? Is it an algorithm, any code related to this? I can not search related files in raft_doc