Off the Beaten Path: Let's Replace Term-Based Retrieval with k-NN Search

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216224%3A14330%2F16%3A00088811" target="_blank" >RIV/00216224:14330/16:00088811 - isvavai.cz</a>

Výsledek na webu

<a href="http://dx.doi.org/10.1145/2983323.2983815" target="_blank" >http://dx.doi.org/10.1145/2983323.2983815</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1145/2983323.2983815" target="_blank" >10.1145/2983323.2983815</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Off the Beaten Path: Let's Replace Term-Based Retrieval with k-NN Search

Popis výsledku v původním jazyce

Retrieval pipelines commonly rely on a term-based search to obtain candidate records, which are subsequently re-ranked. Some candidates are missed by this approach, e.g., due to a vocabulary mismatch. We address this issue by replacing the term-based search with a generic k-NN retrieval algorithm, where a similarity function can take into account subtle term associations. While an exact brute-force k-NN search using this similarity function is slow, we demonstrate that an approximate algorithm can be nearly two orders of magnitude faster at the expense of only a small loss in accuracy. A retrieval pipeline using an approximate k-NN search can be more effective and efficient than the term-based pipeline. This opens up new possibilities for designing effective retrieval pipelines. Our software (including data-generating code) and derivative data based on the Stack Overflow collection is available online.(1)

Název v anglickém jazyce

Off the Beaten Path: Let's Replace Term-Based Retrieval with k-NN Search

Popis výsledku anglicky

Retrieval pipelines commonly rely on a term-based search to obtain candidate records, which are subsequently re-ranked. Some candidates are missed by this approach, e.g., due to a vocabulary mismatch. We address this issue by replacing the term-based search with a generic k-NN retrieval algorithm, where a similarity function can take into account subtle term associations. While an exact brute-force k-NN search using this similarity function is slow, we demonstrate that an approximate algorithm can be nearly two orders of magnitude faster at the expense of only a small loss in accuracy. A retrieval pipeline using an approximate k-NN search can be more effective and efficient than the term-based pipeline. This opens up new possibilities for designing effective retrieval pipelines. Our software (including data-generating code) and derivative data based on the Stack Overflow collection is available online.(1)

Druh

D - Stať ve sborníku

CEP obor

IN - Informatika

OECD FORD obor

—

Projekt

<a href="/cs/project/GBP103%2F12%2FG084" target="_blank" >GBP103/12/G084: Centrum pro multi-modální interpretaci dat velkého rozsahu</a><br>

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

CIKM'16: PROCEEDINGS OF THE 2016 ACM CONFERENCE ON INFORMATION AND KNOWLEDGE MANAGEMENT

ISBN

9781450340731

ISSN

—

e-ISSN

—

Počet stran výsledku

10

Strana od-do

1099-1108

Název nakladatele

ASSOC COMPUTING MACHINERY

Místo vydání

NEW YORK

Místo konání akce

IUPUI, Indianapolis, IN

Datum konání akce

24. 10. 2016

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000390890800113