Language processing and learning models for community question answering in Arabic

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

In this paper we focus on the problem of question ranking in community question answering (cQA) forums in Arabic. We address the task with machine learning algorithms using advanced Arabic text representations. The latter are obtained by applying tree kernels to constituency parse trees combined with textual similarities, including word embeddings. Our two main contributions are: (i)[U+202F]an Arabic language processing pipeline based on UIMA-from segmentation to constituency parsing-built on top of Farasa, a state-of-the-art Arabic language processing toolkit; and (ii)[U+202F]the application of long short-term memory neural networks to identify the best text fragments in questions to be used in our tree-kernel-based ranker. Our thorough experimentation on a recently released cQA dataset shows that the Arabic linguistic processing provided by Farasa produces strong results and that neural networks combined with tree kernels further boost the performance in terms of both efficiency and accuracy. Our approach also enables an implicit comparison between different processing pipelines as our tests on Farasa and Stanford parsers demonstrate.

Original languageEnglish
JournalInformation Processing and Management
DOIs
Publication statusAccepted/In press - 1 Jan 2017

Fingerprint

neural network
Processing
language
learning
community
ranking
Pipelines
Neural networks
linguistics
efficiency
Linguistics
Learning algorithms
performance
Learning systems
Language
Question answering
Learning model
Kernel
segmentation

Keywords

  • Attention models
  • Community question answering
  • Constituency parsing in Arabic
  • Long short-term memory neural networks
  • Tree-kernel-based ranking

ASJC Scopus subject areas

  • Information Systems
  • Media Technology
  • Computer Science Applications
  • Management Science and Operations Research
  • Library and Information Sciences

Cite this

@article{ea88dcaab3824dd1a3f89c6d8ba24692,
title = "Language processing and learning models for community question answering in Arabic",
abstract = "In this paper we focus on the problem of question ranking in community question answering (cQA) forums in Arabic. We address the task with machine learning algorithms using advanced Arabic text representations. The latter are obtained by applying tree kernels to constituency parse trees combined with textual similarities, including word embeddings. Our two main contributions are: (i)[U+202F]an Arabic language processing pipeline based on UIMA-from segmentation to constituency parsing-built on top of Farasa, a state-of-the-art Arabic language processing toolkit; and (ii)[U+202F]the application of long short-term memory neural networks to identify the best text fragments in questions to be used in our tree-kernel-based ranker. Our thorough experimentation on a recently released cQA dataset shows that the Arabic linguistic processing provided by Farasa produces strong results and that neural networks combined with tree kernels further boost the performance in terms of both efficiency and accuracy. Our approach also enables an implicit comparison between different processing pipelines as our tests on Farasa and Stanford parsers demonstrate.",
keywords = "Attention models, Community question answering, Constituency parsing in Arabic, Long short-term memory neural networks, Tree-kernel-based ranking",
author = "Salvatore Romeo and Giovanni Martino and Yonatan Belinkov and Alberto Barron and Mohamed Eldesouki and Kareem Darwish and Hamdy Mubarak and James Glass and Alessandro Moschitti",
year = "2017",
month = "1",
day = "1",
doi = "10.1016/j.ipm.2017.07.003",
language = "English",
journal = "Information Processing and Management",
issn = "0306-4573",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Language processing and learning models for community question answering in Arabic

AU - Romeo, Salvatore

AU - Martino, Giovanni

AU - Belinkov, Yonatan

AU - Barron, Alberto

AU - Eldesouki, Mohamed

AU - Darwish, Kareem

AU - Mubarak, Hamdy

AU - Glass, James

AU - Moschitti, Alessandro

PY - 2017/1/1

Y1 - 2017/1/1

N2 - In this paper we focus on the problem of question ranking in community question answering (cQA) forums in Arabic. We address the task with machine learning algorithms using advanced Arabic text representations. The latter are obtained by applying tree kernels to constituency parse trees combined with textual similarities, including word embeddings. Our two main contributions are: (i)[U+202F]an Arabic language processing pipeline based on UIMA-from segmentation to constituency parsing-built on top of Farasa, a state-of-the-art Arabic language processing toolkit; and (ii)[U+202F]the application of long short-term memory neural networks to identify the best text fragments in questions to be used in our tree-kernel-based ranker. Our thorough experimentation on a recently released cQA dataset shows that the Arabic linguistic processing provided by Farasa produces strong results and that neural networks combined with tree kernels further boost the performance in terms of both efficiency and accuracy. Our approach also enables an implicit comparison between different processing pipelines as our tests on Farasa and Stanford parsers demonstrate.

AB - In this paper we focus on the problem of question ranking in community question answering (cQA) forums in Arabic. We address the task with machine learning algorithms using advanced Arabic text representations. The latter are obtained by applying tree kernels to constituency parse trees combined with textual similarities, including word embeddings. Our two main contributions are: (i)[U+202F]an Arabic language processing pipeline based on UIMA-from segmentation to constituency parsing-built on top of Farasa, a state-of-the-art Arabic language processing toolkit; and (ii)[U+202F]the application of long short-term memory neural networks to identify the best text fragments in questions to be used in our tree-kernel-based ranker. Our thorough experimentation on a recently released cQA dataset shows that the Arabic linguistic processing provided by Farasa produces strong results and that neural networks combined with tree kernels further boost the performance in terms of both efficiency and accuracy. Our approach also enables an implicit comparison between different processing pipelines as our tests on Farasa and Stanford parsers demonstrate.

KW - Attention models

KW - Community question answering

KW - Constituency parsing in Arabic

KW - Long short-term memory neural networks

KW - Tree-kernel-based ranking

UR - http://www.scopus.com/inward/record.url?scp=85027252939&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85027252939&partnerID=8YFLogxK

U2 - 10.1016/j.ipm.2017.07.003

DO - 10.1016/j.ipm.2017.07.003

M3 - Article

JO - Information Processing and Management

JF - Information Processing and Management

SN - 0306-4573

ER -