Fast Single-Core K-Nearest Neighbor Graph Computation

Kluser, Dan; Bokstaller, Jonas; Rutz, Samuel; Buner, Tobias

doi:10.48550/arxiv.2112.06630

Cited by 1 publication

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Parallelization? Termination Dong et al [7] C++ no OpenMP & map-reduce δ-proportion update McInnes [14] Python no none no possible update Kluser et al [12] C no none no possible update This paper Java yes fork-join pool statistical criterion is reported in [7]. Nor shall we outline the different approaches to K-nearest neighbor approximation, briefly surveyed by Aumüller, Bernhardsson and, Faithfull [1].…”

Section: Authorsmentioning

confidence: 99%

See 1 more Smart Citation

Empirical complexity of comparator-based nearest neighbor descent

Baron¹,

Darling²

2022

Preprint

View full text Add to dashboard Cite

A Java parallel streams implementation of the K-nearest neighbor descent algorithm is presented using a natural statistical termination criterion. Input data consist of a set S of n objects of type V, and a Function>, which enables any x ∈ S to decide which of y, z ∈ S \ {x} is more similar to x. Experiments with the Kullback-Leibler divergence Comparator support the prediction that the number of rounds of K-nearest neighbor updates need not exceed twice the diameter of the undirected version of a random regular out-degree K digraph on n vertices. Overall complexity was O(nK 2 log K (n)) in the class of examples studied. When objects are sampled uniformly from a d-dimensional simplex, accuracy of the K-nearest neighbor approximation is high up to d = 20, but declines in higher dimensions, as theory would predict.

show abstract

Section: Authorsmentioning

confidence: 99%

“…Single-threaded C Implementation. Kluser et al [12] describe a runtime-optimized C implementation for the 2 -distance metric, and report performance improvements over the two versions above. Their approach increases locality by improving the otherwise irregular memory access pattern.…”

Section: Python Implementationmentioning

confidence: 99%

Empirical complexity of comparator-based nearest neighbor descent

Baron¹,

Darling²

2022

Preprint

View full text Add to dashboard Cite

show abstract

Fast Single-Core K-Nearest Neighbor Graph Computation

Cited by 1 publication

References 0 publications

Empirical complexity of comparator-based nearest neighbor descent

Empirical complexity of comparator-based nearest neighbor descent

Contact Info

Product

Resources

About