Co-reference resolution is defined as the task of determining which mentions in a discourse refer to the same real world entity. It is important for higher level tasks like question answering, event detection from news, dialog systems, and chat bots. Formally, the input is a text in which some noun phrases are tagged as the entities; and the participating teams should find which ones are co-referred. Similar Tasks: pronoun resolution. This task can be regarded as a special case of co-reference resolution.
There is a Farsi corpus with one million tokens annotated co-referring noun phrases. This corpus is publicly available and is introduced as the training data: http://www.parsigan.ir/projects/Coref . You will have to create an account to be able to access the corpus.
The test data will be prepared according to the timeline of the shared task. Evaluation Procedure There are several evaluation measure for this task: MUC, B3, BLANC, CAEF, and LEA. Each of them have a different definition for precision and recall. In the proposed task, we try to report all of them for the participating teams.
Training data format: The training data contain several documents. Documents are in CoNLL format and there are six columns separated by a tab. Each word has been put on a separate line and there is an empty line after each sentence. The first item on each line is a word; the second specifies its named entity tag; the third one is an id which is unique in that document; the forth one is co-ref id; the fifth one is co-ref type; and the sixth one specifies animate/inanimate feature for the entity.
Input Format: Each word is in a separate line and empty line between sentences.
Output Format:Two columns for word and co-ref id.
In case of any problem in receiving data, please contact email@example.com
A simple baseline will be developed by the task organizer. It looks first for the nearest preceding individual that is compatible with the referring expression
Important Dates: http://nsurl.org/importantdates/
To participate in this task , the team leader has to do the following:
- Choose a name for your team (The name should reflect your team)
- login as an author to https://easychair.org/conferences/?conf=nsurl2019
- add the paper title: Team-name at NSURL-2019 Task 6: Co-reference Resolution in Farsi
- Paper authors of the paper: The team members
- Paper abstract and keywords: add a simple tentative abstract that you can modify anytime
We list here the results of the participating teams after 30 June 2019.
We list here instructions for paper submissions after 30.June 2019.
If you have any queries regarding this task, please refer to the task organizers:
Dr. Heshaam Faili <firstname.lastname@example.org >
Nasrin Taghizadeh < email@example.com >