Explaining Natural Language query results

Daniel Deutch; Nave Frost; Amir Gilad

doi:10.1007/s00778-019-00584-7

Explaining Natural Language query results

Daniel Deutch, Nave Frost, Amir Gilad^*

^*Corresponding author for this work

School of Computer Science

Research output: Contribution to journal › Article › peer-review

Abstract

Multiple lines of research have developed Natural Language (NL) interfaces for formulating database queries. We build upon this work, but focus on presenting a highly detailed form of the answers in NL. The answers that we present are importantly based on the provenance of tuples in the query result, detailing not only the results but also their explanations. We develop a novel method for transforming provenance information to NL, by leveraging the original NL query structure. Furthermore, since provenance information is typically large and complex, we present two solutions for its effective presentation as NL text: one that is based on provenance factorization, with novel desiderata relevant to the NL case and one that is based on summarization. We have implemented our solution in an end-to-end system supporting questions, answers and provenance, all expressed in NL. Our experiments, including a user study, indicate the quality of our solution and its scalability.

Original language	English
Pages (from-to)	485-508
Number of pages	24
Journal	VLDB Journal
Volume	29
Issue number	1
DOIs	https://doi.org/10.1007/s00778-019-00584-7
State	Published - 1 Jan 2020

Keywords

CQ
NL
Natural Language
Provenance
UCQ

Access to Document

10.1007/s00778-019-00584-7

Cite this

@article{ac8cc33957314d53961ce6bc63bc27e5,

title = "Explaining Natural Language query results",

abstract = "Multiple lines of research have developed Natural Language (NL) interfaces for formulating database queries. We build upon this work, but focus on presenting a highly detailed form of the answers in NL. The answers that we present are importantly based on the provenance of tuples in the query result, detailing not only the results but also their explanations. We develop a novel method for transforming provenance information to NL, by leveraging the original NL query structure. Furthermore, since provenance information is typically large and complex, we present two solutions for its effective presentation as NL text: one that is based on provenance factorization, with novel desiderata relevant to the NL case and one that is based on summarization. We have implemented our solution in an end-to-end system supporting questions, answers and provenance, all expressed in NL. Our experiments, including a user study, indicate the quality of our solution and its scalability.",

keywords = "CQ, NL, Natural Language, Provenance, UCQ",

author = "Daniel Deutch and Nave Frost and Amir Gilad",

note = "Publisher Copyright: {\textcopyright} 2019, Springer-Verlag GmbH Germany, part of Springer Nature.",

year = "2020",

month = jan,

day = "1",

doi = "10.1007/s00778-019-00584-7",

language = "אנגלית",

volume = "29",

pages = "485--508",

journal = "VLDB Journal",

issn = "1066-8888",

publisher = "Springer New York",

number = "1",

}

TY - JOUR

T1 - Explaining Natural Language query results

AU - Deutch, Daniel

AU - Frost, Nave

AU - Gilad, Amir

PY - 2020/1/1

Y1 - 2020/1/1

N2 - Multiple lines of research have developed Natural Language (NL) interfaces for formulating database queries. We build upon this work, but focus on presenting a highly detailed form of the answers in NL. The answers that we present are importantly based on the provenance of tuples in the query result, detailing not only the results but also their explanations. We develop a novel method for transforming provenance information to NL, by leveraging the original NL query structure. Furthermore, since provenance information is typically large and complex, we present two solutions for its effective presentation as NL text: one that is based on provenance factorization, with novel desiderata relevant to the NL case and one that is based on summarization. We have implemented our solution in an end-to-end system supporting questions, answers and provenance, all expressed in NL. Our experiments, including a user study, indicate the quality of our solution and its scalability.

AB - Multiple lines of research have developed Natural Language (NL) interfaces for formulating database queries. We build upon this work, but focus on presenting a highly detailed form of the answers in NL. The answers that we present are importantly based on the provenance of tuples in the query result, detailing not only the results but also their explanations. We develop a novel method for transforming provenance information to NL, by leveraging the original NL query structure. Furthermore, since provenance information is typically large and complex, we present two solutions for its effective presentation as NL text: one that is based on provenance factorization, with novel desiderata relevant to the NL case and one that is based on summarization. We have implemented our solution in an end-to-end system supporting questions, answers and provenance, all expressed in NL. Our experiments, including a user study, indicate the quality of our solution and its scalability.

KW - CQ

KW - NL

KW - Natural Language

KW - Provenance

KW - UCQ

UR - http://www.scopus.com/inward/record.url?scp=85075894269&partnerID=8YFLogxK

U2 - 10.1007/s00778-019-00584-7

DO - 10.1007/s00778-019-00584-7

M3 - מאמר

AN - SCOPUS:85075894269

VL - 29

SP - 485

EP - 508

JO - VLDB Journal

JF - VLDB Journal

SN - 1066-8888

IS - 1

ER -

Explaining Natural Language query results

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this