​

​

Publications

2012

The PhD thesis of Amal Alsaif. 2012. Human and Automatic Annotation of Discourse Relations for Arabic. (pdf)

2011

 Al-saif, A. and Markert, K. Modelling Discourse Relations for Arabic. The proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, July 2011, Edinburgh. (pdf)

Al-saif, A. Annotating Discourse Connectives in MSA: Disagreement Cases in the LADTB. Corpus Linguistics 2011, Discourse and Corpus Linguistics, Birmingham, Uk, 2011. (ppt)

2010

Al-Saif, A; Markert, K. 2010. The Leeds Arabic Discourse Treebank: Annotating discourse connectives for Arabic. The Proc. of the conference on Language Resources and Evaluation. Malta.(pdf)

2009

Al-Saif, Markert, and Abdul-Raof. 2009. Corpus-Based Study: Extensive Collection of Discourse Connectives For Arabic. In Proceedings of The Saudi International Conference 2009 (SIC09), Surrey, UK.(pdf)​

READ: the annotation tool for Arabic discourse:

We developed the first discourse annotation tool for Arabic (READ: Relation annotation for English and Arabic Discourse) to ensure a reliable annotation and to response to the Arabic specific  requirments. The annotation in the first version of the tool is a stand-off style (based on the raw texts only) to ensure more flexibility in the READ tool, it can be used to annotate text without its syntactic annotation.

The tool can be used for Arabic and English and any Unicode languages. The interface’s language will be switched accordingly.  

Firstly, the potential discourse connectives in the text are highlighted. The annotator uses the arrows to move the actual discourse connectives into the appropriate list on the left. Not disourse connectives must be moved into the list on the right. Next, for each discourse connective the annotator marks text span of the arguments Arg1 and Arg2 then uses the related buttons.

Finally, the annotator must select one or more discourse relations from the drop-down list. Then, save the annotation before jump into next potential discourse connective in the middle list. 

Downloads:

• The READ Tool package: contains executable jar file plus to a user manual and the potential discourse connectives files for Arabic and English (using the discourse connectives annotated in the PDTB). It is possible to add or remove connectives from the files before run the tool. (zipped_file: READ_Arabic_discourse_annotation_tool).

• To use:
• Our collection of the Arabic discourse connectives with English translation. 

• The annotation guidelines for the discourse relations and connectives in the LADTB.

• Distribution of discourse connectives and associated relations and connectives in the LADTB.
• 
  
• Please see the appendices of the thesis Human and Automatic Discourse Annotation for Relations in Arabic