[Corpora-List] Crowdsourcing for arabic tasks

Noura Farra noura at cs.columbia.edu
Sat Aug 8 02:26:00 CEST 2015


Hi Imen,

I have done some work on sentiment annotation for Arabic using Amazon Mechanical Turk. I was able to find a reasonable number of workers on AMT who are fluent in Arabic. I ended up assigning the HITs to a qualified group of workers who were really good at the sentiment task.

For your reference here is the link to the paper:

Annotating Targets of Sentiment in Arabic Using Crowdsourcing

http://www.aclweb.org/anthology/W/W15/W15-32.pdf#page=101

Best, Noura

On Fri, Aug 7, 2015 at 10:38 AM, Mahmoud EL-Haj <dr.melhaj at gmail.com> wrote:


> Dear Imen,
>
>
>
> We have done some work on using AMT/Crowdflower to summarise Arabic
> documents [1] and also using experts to build Arabic resources needed for
> MultiLing 2011/2013 tasks [2],[3]. In the meanwhile we are working on
> hiring Arabic native speaker experts to build an Arabic lexicon.
>
> You should be able to run your tasks on AMT but you need to be careful
> when it comes to evaluating the quality and consistency of the results.
>
>
>
> References:
>
>
>
> [1] Using Mechanical Turk to Create a Corpus of Arabic Summaries:
>
> http://www.lancaster.ac.uk/staff/elhaj/docs/LREC2010-MTurk-Final_v2.pdf
>
>
>
> [2] TAC 2011 MultiLing Pilot Overview
>
>
> http://www.nist.gov/tac/publications/2011/additional.papers/Summarization2011_MultiLing_overview.proceedings.pdf
>
>
>
> [3] Multi-document multilingual summarization corpus preparation, Part 1:
> Arabic, English, Greek, Chinese, Romanian
>
> http://aclweb.org/anthology/W/W13/W13-3101.pdf
>
>
>
> Best,
> Mahmoud
>
>
>
> --
>
> Dr Mahmoud El-Haj
>
> Senior Research Associate
>
> School of Computing and Communications
>
> Lancaster University
>
> http://www.lancaster.ac.uk/staff/elhaj/
>
> m.el-haj at lancaster.ac.uk
>
>
>
>
>
>
>
>
>
> *From:* corpora-bounces at uib.no [mailto:corpora-bounces at uib.no] *On Behalf
> Of *imen touati
> *Sent:* Wednesday, August 05, 2015 9:10 PM
> *To:* corpora at uib.no
> *Subject:* [Corpora-List] Crowdsourcing for arabic tasks
>
>
>
> Dear all,
>
>
>
> Please have any one idea about any group like Amazon Mechanical turk that
> doing crowdsourcing for tasks in arabic and that employs arabic language
> expert (more precisely the task is the annotation of arabic documents in
> order to do opinion mining and sentiment analysis ).
>
>
>
> Thank you in advance.
>
>
>
>
>
>
> *....................................................Imen Touati*
>
>
> *PhD Computer Science student*
>
> *Faculty of Economic Sciences and management of Sfax*
> *ANLP Research Group*
> http://sites.google.com/site/anlprg
>
> *MIRACL Laboratory*
> www.miracl.rnu.tn
>
> *Address :* *FSEGS, BP 1088, 3018 Sfax, Tunisia*
> Email : i <bayoudhi.amine at gmail.com>smi_touati at yahoo.fr
>
> _______________________________________________
> UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
> Corpora mailing list
> Corpora at uib.no
> http://mailman.uib.no/listinfo/corpora
>
>
-------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/html Size: 11928 bytes Desc: not available URL: <https://mailman.uib.no/public/corpora/attachments/20150807/5e4bbb05/attachment.txt>



More information about the Corpora mailing list