PPP-Codes for Large-Scale Similarity Searching

Kód výsledku v IS VaVaI

RIV/00216224:14330/16:00087959 - isvavai.cz

Výsledek na webu

http://link.springer.com/chapter/10.1007/978-3-662-49214-7_2

DOI - Digital Object Identifier

10.1007/978-3-662-49214-7_2

Jazyk výsledku

angličtina

Název v původním jazyce

PPP-Codes for Large-Scale Similarity Searching

Popis výsledku v původním jazyce

Many current applications need to organize data with respect to mutual similarity between data objects. A typical general strategy to retrieve objects similar to a given sample is to access and then refine a candidate set of objects. We propose an indexing and search technique that can significantly reduce the candidate set size by combination of several space partitionings. Specifically, we propose a mapping of objects from a generic metric space onto main memory codes using several pivot spaces; our search algorithm first ranks objects within each pivot space and then aggregates these rankings producing a candidate set reduced by two orders of magnitude while keeping the same answer quality. Our approach is designed to well exploit contemporary HW: (1) larger main memories allow us to use rich and fast index, (2) multi-core CPUs well suit our parallel search algorithm, and (3) SSD disks without mechanical seeks enable efficient selective retrieval of candidate objects.

Název v anglickém jazyce

PPP-Codes for Large-Scale Similarity Searching

Popis výsledku anglicky

Many current applications need to organize data with respect to mutual similarity between data objects. A typical general strategy to retrieve objects similar to a given sample is to access and then refine a candidate set of objects. We propose an indexing and search technique that can significantly reduce the candidate set size by combination of several space partitionings. Specifically, we propose a mapping of objects from a generic metric space onto main memory codes using several pivot spaces; our search algorithm first ranks objects within each pivot space and then aggregates these rankings producing a candidate set reduced by two orders of magnitude while keeping the same answer quality. Our approach is designed to well exploit contemporary HW: (1) larger main memories allow us to use rich and fast index, (2) multi-core CPUs well suit our parallel search algorithm, and (3) SSD disks without mechanical seeks enable efficient selective retrieval of candidate objects.

Druh

D - Stať ve sborníku

CEP obor

IN - Informatika

OECD FORD obor

—

Projekt

GBP103/12/G084: Centrum pro multi-modální interpretaci dat velkého rozsahu

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2016

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Transactions on Large-Scale Data- and Knowledge-Centered Systems XXIV

ISBN

9783662492130

ISSN

0302-9743

e-ISSN

—

Počet stran výsledku

27

Strana od-do

61-87

Název nakladatele

Springer

Místo vydání

Berlin Heidelberg

Místo konání akce

Berlin Heidelberg

Datum konání akce

1. 1. 2016

Typ akce podle státní příslušnosti

CST - Celostátní akce

Kód UT WoS článku

—