AbjadNLP 2027 · Shared task

AbjadAuthorID 2027

Authorship identification in Kurdish and Urdu

Identify the author of a literary passage from a set of candidate authors. The proposed 2027 edition focuses on Kurdish and Urdu, languages written using Arabic-derived scripts.

Explore the task ↓Participation details

Language focus

Kurdish (Sorani script)زمانی کوردی
Urduاردو زبان

Illustrative language labels, not dataset samples.

2027 proposed scope: Kurdish and Urdu. MSA is not currently confirmed.

About the shared task

Who wrote this passage?

AbjadAuthorID is a multiclass authorship attribution task. Given a literary text excerpt, participating systems identify its author from a defined set of candidate authors. The 2027 edition proposes Kurdish and Urdu tracks, building on the previous task, which also included Modern Standard Arabic.

Input

Literary text excerpt

A paragraph or other passage drawn from literary writing, with an associated candidate-author set.

Output

Predicted author

A single author label or identifier selected from the task's candidate authors.

Proposed track

Kurdish

Authorship identification for Kurdish texts written in an Arabic-derived script. Dialect and corpus coverage will be confirmed by the organisers.

Proposed track

Urdu

Authorship identification for Urdu literary texts, using linguistic and stylistic evidence to distinguish writers.

Data and evaluation

The previous edition described literary excerpts from novels, short stories and poetry, with macro-F1 as its primary metric and accuracy as a secondary metric. The 2027 training and test sets, licensing, final tracks and evaluation protocol remain to be confirmed.

Taking part

Participation details

Registration, data access, submission instructions and 2027 deadlines will be announced by the organising team.

For background, see the previous AbjadAuthorID website. Its 2026 dates, platforms and MSA track do not automatically carry over to 2027.

Visit the AbjadNLP 2027 workshop page →
Proposed languages
Kurdish and Urdu
MSA track
Not confirmed
Data access and licensing
To be announced
Evaluation protocol
To be confirmed
Registration and submissions
To be announced
Shared-task dates
To be announced

The team

Shared-task organisers

Organisers listed on the previous edition’s website. The 2027 organising team is subject to confirmation.

  • Shadi AbudalfaKing Fahd University of Petroleum & Minerals
  • Saad EzziniKing Fahd University of Petroleum & Minerals
  • Ahmed AbdelaliHUMAIN
  • Mustafa JarrarHamad Bin Khalifa University / Birzeit University
  • Mo El-HajVinUniversity
  • Nadir DurraniHamad Bin Khalifa University
  • Hassan SajjadDalhousie University
  • Farah AdeebaUniversity of Massachusetts Amherst
  • Sina AhmadiUniversity of Zurich