Task Description:

Speech acts are the type of communicative acts within a conversation.  Speech act recognition (aka classification) has been an active research in recent years. However, much less attention was directed towards this task in Arabic due to the lack of resources for training an Arabic speech-act classifier.  We utilized the recently published dataset called Arabic Speech Act and Sentiment (ArSAS) corpus for Arabic tweets by (Elmadany et al, 2108).

Data ArSAS, a large set of 21k Arabic tweets covering multiple topics were collected, prepared and annotated or six different classes of speech-act labels, such as expression, assertion, and question.  In addition, the same set of tweets were also annotated with four classes of sentiment.

ArSAS tweets are collected from three main types of 20 topics: Long-standing (topics about articles that are commonly discussed over long period of time), Entity (topics about celebrities or organization) and Event (topics about an important thing that is happening).

ArSAS corpus six speech act tags:

• Assertion: user declare some preposition such as stating, claiming, reporting or announcing
• Recommendation: user recommending something
• Expression: user expresses some psychological state such as thanking, apologizing or congratulating
• Question: user asks a question such as why, what or confirmation
• Request: user asks for something such as ordering, requesting or demanding.
• Miscellaneous: user committed to some future action such as promising or offering.

For sentiment analysis tags, positive, negative, mixed(contains  both  positive and negative sentiment),  or neutral(no opinion or sentiment disclosed).

ArSAS released dataset contains the following information:

• ID: ID of the tweet.
• Text: the original unprocessed text of the tweet
• Topic: topic type of the keyword used to collect the tweet.
• Sentiment: selected sentiment label.
• Sentiment Conf.: Confidence score of sentiment label.
• Speech Act: selected speech act label.
• Speech Act Conf.: Confidence   score   of   the speech act label.

For more information about the distribution of speech acts tags and sentiment analysis tags in ArSAS, kindly read this paper [Elmadany, A., Mubarak, H., & Magdy, W. (2018). Arsas: An arabic speech-act and sentiment corpus of tweets. Proceedings of the 3rd workshop on open-source Arabic corpora and processing tools.]

Sub tasks:

This task has two subtasks: (5.1) sentiment analysis classification (5.2) speech acts classification. Moreover, we need to study the effectiveness of using sentiment tags in speech act recognition.

Task (5.1) sentiment analysis classification:

Input: tweet text

Output: sentiment tags

Task (5.2) speech act classification:

Input: tweet text

Output: speech act tags

Tweets Samples:

Data:

We used ArSAS data set. Train set and test set data will be announced soon.

Important Dates: http://nsurl.org/importantdates

Task participation:

To participate in this task, the team leader has to do the following:

1. Choose a name for your team (The name should reflect your team)
2. login as an author to https://easychair.org/conferences/?conf=nsurl2019
3. add the paper title: <Team-name> at NSURL-2019 Task 5: Arabic Speech-Act and Sentiment Classification.
4. Paper authors of the paper: The team members
5. Paper abstract and keywords: add a simple tentative abstract that you can modify anytime
6. submit

Results:

We list here the results of the participating teams after 30 June 2019.

Paper submission:

We list here instructions for paper submissions after 30.June 2019.

Task Organizers:

If you have any queries regarding this task, please refer to the task organizers:

AbdelRahim Elmadany: <aelmadany@jazanu.edu.sa>

Hamdy Mubarak: < hmubarak@hbku.edu.qa>

Abed Alhakim Freihat: <abed.freihat@unitn.it>