@inproceedings{antoun-etal-2021-aragpt2,
title = "{A}ra{GPT}2: Pre-Trained Transformer for {A}rabic Language Generation",
author = "Antoun, Wissam and
Baly, Fady and
Hajj, Hazem",
editor = "Habash, Nizar and
Bouamor, Houda and
Hajj, Hazem and
Magdy, Walid and
Zaghouani, Wajdi and
Bougares, Fethi and
Tomeh, Nadi and
Abu Farha, Ibrahim and
Touileb, Samia",
booktitle = "Proceedings of the Sixth Arabic Natural Language Processing Workshop",
month = apr,
year = "2021",
address = "Kyiv, Ukraine (Virtual)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.wanlp-1.21",
pages = "196--207",
abstract = "Recently, pre-trained transformer-based architectures have proven to be very efficient at language modeling and understanding, given that they are trained on a large enough corpus. Applications in language generation for Arabic are still lagging in comparison to other NLP advances primarily due to the lack of advanced Arabic language generation models. In this paper, we develop the first advanced Arabic language generation model, AraGPT2, trained from scratch on a large Arabic corpus of internet text and news articles. Our largest model, AraGPT2-mega, has 1.46 billion parameters, which makes it the largest Arabic language model available. The mega model was evaluated and showed success on different tasks including synthetic news generation, and zero-shot question answering. For text generation, our best model achieves a perplexity of 29.8 on held-out Wikipedia articles. A study conducted with human evaluators showed the significant success of AraGPT2-mega in generating news articles that are difficult to distinguish from articles written by humans. We thus develop and release an automatic discriminator model with a 98{\%} percent accuracy in detecting model-generated text. The models are also publicly available, hoping to encourage new research directions and applications for Arabic NLP.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="antoun-etal-2021-aragpt2">
<titleInfo>
<title>AraGPT2: Pre-Trained Transformer for Arabic Language Generation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wissam</namePart>
<namePart type="family">Antoun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fady</namePart>
<namePart type="family">Baly</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hazem</namePart>
<namePart type="family">Hajj</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2021-04</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Sixth Arabic Natural Language Processing Workshop</title>
</titleInfo>
<name type="personal">
<namePart type="given">Nizar</namePart>
<namePart type="family">Habash</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Houda</namePart>
<namePart type="family">Bouamor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hazem</namePart>
<namePart type="family">Hajj</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Walid</namePart>
<namePart type="family">Magdy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wajdi</namePart>
<namePart type="family">Zaghouani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fethi</namePart>
<namePart type="family">Bougares</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nadi</namePart>
<namePart type="family">Tomeh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ibrahim</namePart>
<namePart type="family">Abu Farha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Samia</namePart>
<namePart type="family">Touileb</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Kyiv, Ukraine (Virtual)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Recently, pre-trained transformer-based architectures have proven to be very efficient at language modeling and understanding, given that they are trained on a large enough corpus. Applications in language generation for Arabic are still lagging in comparison to other NLP advances primarily due to the lack of advanced Arabic language generation models. In this paper, we develop the first advanced Arabic language generation model, AraGPT2, trained from scratch on a large Arabic corpus of internet text and news articles. Our largest model, AraGPT2-mega, has 1.46 billion parameters, which makes it the largest Arabic language model available. The mega model was evaluated and showed success on different tasks including synthetic news generation, and zero-shot question answering. For text generation, our best model achieves a perplexity of 29.8 on held-out Wikipedia articles. A study conducted with human evaluators showed the significant success of AraGPT2-mega in generating news articles that are difficult to distinguish from articles written by humans. We thus develop and release an automatic discriminator model with a 98% percent accuracy in detecting model-generated text. The models are also publicly available, hoping to encourage new research directions and applications for Arabic NLP.</abstract>
<identifier type="citekey">antoun-etal-2021-aragpt2</identifier>
<location>
<url>https://aclanthology.org/2021.wanlp-1.21</url>
</location>
<part>
<date>2021-04</date>
<extent unit="page">
<start>196</start>
<end>207</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T AraGPT2: Pre-Trained Transformer for Arabic Language Generation
%A Antoun, Wissam
%A Baly, Fady
%A Hajj, Hazem
%Y Habash, Nizar
%Y Bouamor, Houda
%Y Hajj, Hazem
%Y Magdy, Walid
%Y Zaghouani, Wajdi
%Y Bougares, Fethi
%Y Tomeh, Nadi
%Y Abu Farha, Ibrahim
%Y Touileb, Samia
%S Proceedings of the Sixth Arabic Natural Language Processing Workshop
%D 2021
%8 April
%I Association for Computational Linguistics
%C Kyiv, Ukraine (Virtual)
%F antoun-etal-2021-aragpt2
%X Recently, pre-trained transformer-based architectures have proven to be very efficient at language modeling and understanding, given that they are trained on a large enough corpus. Applications in language generation for Arabic are still lagging in comparison to other NLP advances primarily due to the lack of advanced Arabic language generation models. In this paper, we develop the first advanced Arabic language generation model, AraGPT2, trained from scratch on a large Arabic corpus of internet text and news articles. Our largest model, AraGPT2-mega, has 1.46 billion parameters, which makes it the largest Arabic language model available. The mega model was evaluated and showed success on different tasks including synthetic news generation, and zero-shot question answering. For text generation, our best model achieves a perplexity of 29.8 on held-out Wikipedia articles. A study conducted with human evaluators showed the significant success of AraGPT2-mega in generating news articles that are difficult to distinguish from articles written by humans. We thus develop and release an automatic discriminator model with a 98% percent accuracy in detecting model-generated text. The models are also publicly available, hoping to encourage new research directions and applications for Arabic NLP.
%U https://aclanthology.org/2021.wanlp-1.21
%P 196-207
Markdown (Informal)
[AraGPT2: Pre-Trained Transformer for Arabic Language Generation](https://aclanthology.org/2021.wanlp-1.21) (Antoun et al., WANLP 2021)
ACL