Smaller, Faster & Lighter KNN Graph Constructions - Inria EPFL Access content directly
Conference Papers Year : 2020

Smaller, Faster & Lighter KNN Graph Constructions

Abstract

We propose GoldFinger, a new compact and fast-to-compute binary representation of datasets to approximate Jaccard's index. We illustrate the effectiveness of GoldFinger on the emblematic big data problem of K-Nearest-Neighbor (KNN) graph construction and show that GoldFinger can drastically accelerate a large range of existing KNN algorithms with little to no overhead. As a side effect, we also show that the compact representation of the data protects users' privacy for free by providing k-anonymity and l-diversity. Our extensive evaluation of the resulting approach on several realistic datasets shows that our approach delivers speedups of up to 78.9% compared to the use of raw data while only incurring a negligible to moderate loss in terms of KNN quality. To convey the practical value of such a scheme, we apply it to item recommendation and show that the loss in recommendation quality is negligible.
Fichier principal
Vignette du fichier
papier.pdf (1.85 Mo) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-02888286 , version 1 (02-07-2020)

Identifiers

Cite

Rachid Guerraoui, Anne-Marie Kermarrec, Olivier Ruas, François Taïani. Smaller, Faster & Lighter KNN Graph Constructions. WWW '20 - The Web Conference 2020, Apr 2020, Taipei Taiwan, France. pp.1060-1070, ⟨10.1145/3366423.3380184⟩. ⟨hal-02888286⟩
96 View
221 Download

Altmetric

Share

Gmail Facebook X LinkedIn More