Melola Discovery

← Melola Discovery

Open access Jurnal SINTA 1

Atar: Attention-based LSTM for Arabizi transliteration

International Journal of Electrical and Computer Engineering (IJECE) · Vol. 11 No. 3 (2021)

DOI
10.11591/ijece.v11i3.pp2327-2334
ISSN
2722-2578, 2088-8708
Published
2021-06-01
Publisher
Institute of Advanced Engineering and Science
License
creativecommons.org/licenses/by-sa/4.0

Abstract

A non-standard romanization of Arabic script, known as Arbizi, is widely used in Arabic online and SMS/chat communities. However, since state-of-the-art tools and applications for Arabic NLP expects Arabic to be written in Arabic script, handling contents written in Arabizi requires a special attention either by building customized tools or by transliterating them into Arabic script. The latter approach is the more common one and this work presents two significant contributions in this direction. The first one is to collect and publicly release the first large-scale “Arabizi to Arabic script” parallel corpus focusing on the Jordanian dialect and consisting of more than 25 k pairs carefully created and inspected by native speakers to ensure highest quality. Second, we present Atar, an attention-based encoder-decoder model for Arabizi transliteration. Training and testing this model on our dataset yields impressive accuracy (79%) and BLEU score (88.49).

Collaboration network of the authors

17 authors · 47 co-authorship links · 6 across different organisations

Hover or tap a name to highlight its co-authors. Double-click to open the author's page.

  • Jordan University of Science and Technology
  • University of Southampton
  • Yarmouk University
  • Affiliation unknown
  • Different organisations
  • Same organisation
  • Affiliation unknown
  • Thicker line = more joint articles · bigger circle = more articles
Show as table
Co-authorOrganisationJoint articles
Ghadeer Al-BdourJordan University of Science and Technology1
Raffi Al-QurranJordan University of Science and Technology1
Ali M. ShatnawiJordan University of Science and Technology1
Asma’a Al-MnayyisYarmouk University1
Sanaa Abu AlasalJordan University of Science and Technology1
Mohammad A. AlsmiratJordan University of Science and Technology1
Qanita Bani BakerJordan University of Science and Technology1
Amjad NuseirJordan University of Science and Technology1
Hasan AlbalasAffiliation unknown1
Aya NuseirJordan University of Science and Technology1
Maulla AlaliJordan University of Science and Technology1
Firas ZoubiAffiliation unknown1
Mohammed MahdiJordan University of Science and Technology1
Ahmad Al OmariJordan University of Science and Technology1

Affiliations come from OpenAlex; journal OAI-PMH metadata carries none. Authors without an OpenAlex ID are matched by name, which can merge or split people.

This record comes from Crossref (what the publisher deposited), because the journal's own OAI-PMH endpoint could not be harvested.

Metadata source: Crossref API (cadangan, berdasarkan ISSN) https://api.crossref.org/journals/2088-8708/works, first harvested 27 Okt 2026, last seen 27 Okt 2026.
Metadata as oai_dc, converted from Crossref
<oai_dc:dc xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/"><dc:title>Atar: Attention-based LSTM for Arabizi transliteration</dc:title><dc:creator>Talafha, Bashar</dc:creator><dc:creator>Abuammar, Analle</dc:creator><dc:creator>Al-Ayyoub, Mahmoud</dc:creator><dc:description>&lt;jats:p&gt;A non-standard romanization of Arabic script, known as Arbizi, is widely used in Arabic online and SMS/chat communities. However, since state-of-the-art tools and applications for Arabic NLP expects Arabic to be written in Arabic script, handling contents written in Arabizi requires a special attention either by building customized tools or by transliterating them into Arabic script. The latter approach is the more common one and this work presents two significant contributions in this direction. The first one is to collect and publicly release the first large-scale “Arabizi to Arabic script” parallel corpus focusing on the Jordanian dialect and consisting of more than 25 k pairs carefully created and inspected by native speakers to ensure highest quality. Second, we present Atar, an attention-based encoder-decoder model for Arabizi transliteration. Training and testing this model on our dataset yields impressive accuracy (79%) and BLEU score (88.49).&lt;/jats:p&gt;</dc:description><dc:publisher>Institute of Advanced Engineering and Science</dc:publisher><dc:date>2021-06-01</dc:date><dc:type>info:eu-repo/semantics/article</dc:type><dc:format>application/pdf</dc:format><dc:identifier>https://doi.org/10.11591/ijece.v11i3.pp2327-2334</dc:identifier><dc:identifier>10.11591/ijece.v11i3.pp2327-2334</dc:identifier><dc:source>International Journal of Electrical and Computer Engineering (IJECE); Vol 11 No 3 (2021); 2327</dc:source><dc:source>2722-2578</dc:source><dc:source>2088-8708</dc:source><dc:relation>http://ijece.iaescore.com/index.php/IJECE/article/viewFile/22767/14781</dc:relation><dc:rights>http://creativecommons.org/licenses/by-sa/4.0</dc:rights></oai_dc:dc>