IndoNLPWorkshop 2025

@Coling 2025
January 20, 2025 Abu Dhabi, UAE

Shared Task

Task Overview

Typing Romanized Indo-Aryan languages using adhoc transliterations with or without vowels and achieving accurate native script output is a complex and often inaccurate process with the current keyboard systems. This task aims to introduce and evaluate a real-time reverse transliterator that converts Romanized Indo-Aryan languages (Native Language written using English Alphabets) to their native scripts, improving the typing experience for users of these languages.

Objectives of Shared Task

The objectives of the shared task are to develop a real-time transliterator, handle variations, and enhance typing accuracy. The transliterator should be able to handle Ad-hoc Transliterations (short typing scripts) : different typing patterns with or without vowel combinations.

Important Dates

• - 1^st Call for Registration :July 20, 2024
• - 2^nd Call for Registration : July 31, 2024
• - Registration Closing :August 25, 2024 August 31, 2024
• - Intial Briefing :Will be notified via email
• - Interim Progress Meeting I : October 7, 2024
• - Shared Task Completion :TBA

Data

• Dhakshine Dataset
• Aksharantar Dataset
• SwaBhasha Dataset

Submission

- The developed transliteration model/code.
- A report detailing the approach, challenges faced, and solutions implemented.
- A detailed discussion about the evaluation techniques/frameworks used.
- A short paper (4 pages)

Note: All text data and code must be shared under an Apache/MIT style license.

Click to Register :