Efficient Autotuning of Hyperparameters in Approximate Nearest Neighbor Search

Visa fullständig post



Permalänk

http://hdl.handle.net/10138/303722

Citation

Jääsaari , E , Hyvönen , V & Roos , T 2019 , Efficient Autotuning of Hyperparameters in Approximate Nearest Neighbor Search . in Advances in Knowledge Discovery and Data Mining : 23rd Pacific-Asia Conference, PAKDD 2019, Macau, China, April 14-17, 2019, Proceedings, Part II . Lecture Notes in Artificial Intelligence , vol. 11440 , Springer , Cham , pp. 590-602 , Pacific-Asia Conference on Knowledge Discovery and Data Mining 2019 , Macao , China , 14/04/2019 . https://doi.org/10.1007/978-3-030-16145-3_46

Titel: Efficient Autotuning of Hyperparameters in Approximate Nearest Neighbor Search
Författare: Jääsaari, Elias; Hyvönen, Ville; Roos, Teemu
Upphovmannens organisation: Department of Computer Science
Helsinki Institute for Information Technology
Information, Complexity and Learning research group / Teemu Roos
Complex Systems Computation Group
Utgivare: Springer
Datum: 2019-04-17
Språk: eng
Sidantal: 13
Tillhör serie: Advances in Knowledge Discovery and Data Mining
Tillhör serie: Lecture Notes in Artificial Intelligence
ISBN: 978-3-030-16144-6
978-3-030-16145-3
ISSN: 0302-9743
DOI: https://doi.org/10.1007/978-3-030-16145-3_46
Permanenta länken (URI): http://hdl.handle.net/10138/303722
Abstrakt: Approximate nearest neighbor algorithms are used to speed up nearest neighbor search in a wide array of applications. However, current indexing methods feature several hyperparameters that need to be tuned to reach an acceptable accuracy–speed trade-off. A grid search in the parameter space is often impractically slow due to a time-consuming index-building procedure. Therefore, we propose an algorithm for automatically tuning the hyperparameters of indexing methods based on randomized space-partitioning trees. In particular, we present results using randomized k-d trees, random projection trees and randomized PCA trees. The tuning algorithm adds minimal overhead to the index-building process but is able to find the optimal hyperparameters accurately. We demonstrate that the algorithm is significantly faster than existing approaches, and that the indexing methods used are competitive with the state-of-the-art methods in query time while being faster to build.
Subject: 113 Computer and information sciences
Referentgranskad: Ja
Användningsbegränsning: openAccess
Parallelpublicerad version: acceptedVersion


Filer under denna titel

Totalt antal nerladdningar: Laddar...

Filer Storlek Format Granska
Jaasaari_Hyvonen_2019_Efficient_Autotuning.pdf 402.4Kb PDF Granska/Öppna

Detta dokument registreras i samling:

Visa fullständig post