Camel Treebank: An Open Multi-genre Arabic Dependency Treebank

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 2672–2681 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 2672 Camel Treebank: An Open Multi-genre Arabic Dependency Treebank Nizar Habash, Muhammed AbuOdeh, Dima Taji,† Reem Faraj,‡ Jamila El Gizuli,∗ and Omar Kallas Computational Approaches to Modeling Language (CAMeL) Lab New York University Abu Dhabi †Birzeit University ‡Columbia University & CUNY Graduate Center ∗Georgia Institute of Technology nizar.habash@nyu.edu Abstract We present the Camel Treebank (CAMELTB), a 188K word open-source dependency treebank of Modern Standard and Classical Arabic. CAMELTB 1.0 includes 13 sub-corpora comprising selections of texts from pre-Islamic poetry to social media online commentaries, and covering a range of genres from religious and philosophical texts to news, novels, and student essays. The texts are all publicly available (out of copyright, creative commons, or under open licenses). The texts were morphologically tokenized and syntactically parsed automatically, and then manually corrected by a team of trained annotators. The annotations follow the guidelines of the Columbia Arabic Treebank (CATiB) dependency representation. We discuss our annotation process and guideline extensions, and we present some initial observations on lexical and syntactic differences among the annotated sub-corpora. This corpus will be publicly available to support and encourage research on Arabic NLP in general and on new, previously unexplored genres that are of interest to a wider spectrum of researchers, from historical linguistics and digital humanities to computer-assisted language pedagogy. Keywords: Arabic, Syntactic Dependency Treebank, Multiple Genres, Open Source 1. Introduction A lot of research and system development in natural lan- guage processing (NLP) relies heavily on the existence of data enriched with annotations that represent the spe- cific linguistic features of the text. In the case of Arabic, a morphologically rich and complex language, the cre- ation of the Penn Arabic Treebank (PATB) (Maamouri et al., 2004) has led to the creation of many enabling tech- nologies: tokenization, POS tagging, base phrase chunk- ing and syntactic parsing (Pasha et al., 2014; Shahrour et al., 2016; Obeid et al., 2020), among others. Since the creation of the PATB, a number of other tree- banks for Arabic were created with different texts (as opposed to same texts, but different formalisms), e.g., the Columbia Arabic Treebank (CATiB) (Habash and Roth, 2009), the Quran Corpus (Dukes and Buckwalter, 2010), the Arabic Basic Traveling Expressions Corpus (Taji et al., 2018), and most recently the Arabic Poetry Treebank (ArPoT) (Al-Ghamdi et al., 2021). Each of these efforts targets a specific genre, with MSA news getting the lion’s share of attention. The news genre comes with an additional problem, namely, copyright restrictions on the original text, which limits access to news-based treebanks. In this paper we present the Camel Treebank (CAMELTB) a manually annotated large (∼188K words, ∼242K tokens) open-source dependency treebank of Modern Standard Arabic (MSA) and Classical Arabic (CA) in the style of CATiB dependencies. We designed CAMELTB to include selections of texts ranging from pre-Islamic poetry to social media online commentaries, and covers …

Content truncated to 500 words