[open-linguistics] CLEF 2019 Lab ProtestNews Call for Participation

Arda Akdemir aakdemir at ku.edu.tr
Mon Feb 18 00:18:10 UTC 2019

Apologies for cross-posting & please share with anybody who may be
interested in this call.


CLEF ProtestNews 2019

Extracting Protests from News Using Automated Methods


ProtestNews 2019 is a lab that is being organized as part of the Conference
and Labs of the Evaluation Forum 2019 (CLEF). The lab aims at extracting
protest event information from news articles across multiple countries
using machine learning (ML) and generalizable natural language processing
(NLP). This is a sub-task of our ERC-funded research, which aims to
generate a comparative protest databases for China, India, South Africa,
Mexico and Brazil (emw.ku.edu.tr). The lab aims at facilitating research in
line of improving generalizability of natural language processing systems.
Our objective is to develop text classification and information extraction
tools on one country and test them on data from different countries.

The results of our task will appear in the working notes proceedings,

published by CEUR Workshop Proceedings (CEUR-WS.org) and are presented in

the CLEF 2019. Time & Place:  09-12 September 2019, Lugano - Switzerland

We offer three subtasks and target research areas involve but not limited

-    Information Retrieval

-    Natural Language Processing

-    Machine Learning, Deep Learning

-    Big Data, Data Mining

with a focus on generalizability.


We split the ProtestNews task in three subtasks:

1) Task 1: News article classification as protest vs. non-protest is a
binary classification task that aims at discriminating between protest
event related news articles and any other news article. Task 1 forms the
basis of the subsequent tasks and is the first step towards extracting
protest  event related information from news articles.

2) Task 2: Event sentence detection aims at determining event sentences
that contain an event trigger or a mention of it. Event triggers refer to
the textual unit that makes a sentence to be annotated as an event sentence
and an event mention is any kind of mention to that event trigger. For
example; "In the attack that happened yesterday 20 people were injured."
and "Today, we received news that 2 people died in the hospital after the
incident." would be positive examples.

3) Task 3: Event information extraction is an event information extraction
task that targets mainly events. Information such as participants, time and
place of the event. So this task can be considered as a challenging Named
Entity Recognition task where the participants are expected to detect not
only the named entities found in the text but also be able to detect which
ones are related to event. F1 measure will be used to evaluate Task 3.

Subtasks two and three will be based on news articles that are labelled as
protest or not for subtask 1. Participants can choose to participate in one
or more of these subtasks independent of each other. The data from India
will be used for training and testing. Evaluations will be done using the
dataset from China.


Please register before April 26, 2019 (

Data release: March 15, 2019.

Submission deadline is May 10, 2019.


Ali Hürriyetoglu: ahurriyetoglu at ku.edu.tr

Deniz Yüret: dyuret at ku.edu.tr

Erdem Yörük: eryoruk at ku.edu.tr

Çağrı Yoltar: cyoltar at ku.edu.tr

Burak Gürel: bgurel at ku.edu.tr

Fırat Duruşan: fdurusan at ku.edu.tr

Osman Mutlu: omutlu at ku.edu.tr

Arda Akdemir: aakdemir at ku.edu.tr

Theresa Gessler: Theresa.Gessler at eui.eu

Peter Makarov: makarov at cl.uzh.ch


Aline Villavicencio

Antal van den Bosch <http://antalvandenbosch.ruhosting.nl/>

Arzucan Özgür <https://www.cmpe.boun.edu.tr/~ozgur/>

Hristo Tanev

Kemal Oflazer <https://www.andrew.cmu.edu/user/ko/>

Sophia Ananiadou <http://www.nactem.ac.uk/staff/sophia.ananiadou/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/open-linguistics/attachments/20190218/9be2f1f2/attachment-0001.html>

More information about the open-linguistics mailing list