[Apologies for multiple postings]
************************************************************************************
[CFP] CMIR: Code Mixed Information Retrieval From Social Media Data in Bengali-English at FIRE 2025
************************************************************************************
https://cmir-iitbhu.github.io/cmir/index.html
We are excited to announce the Code-Mixed Information Retrieval (CMIR) Shared Task, as part of the Forum for Information Retrieval Evaluation (FIRE 2025), hosted by IIT (BHU) Varanasi from December 17–20, 2025.
Organized by the Information Retrieval Lab (IReL), this shared task addresses the growing challenges of information retrieval in code-mixed social media environments, particularly focusing on Bengali-English conversations.
🎯 Task Overview
Code-mixing, where speakers blend elements of multiple languages within a single sentence, is a common phenomenon in multilingual societies. In India, this often involves using Roman script for Indic languages on social media platforms.
The CMIR task aims to develop robust mechanisms for retrieving the most relevant responses from code-mixed user conversations in social media settings. We provide a code-mixed dataset sourced from community-driven discussions, emphasizing the need for culturally and linguistically informed retrieval techniques.
📅 Important Dates
📂 Training Data Release: June 5, 2025
🧪 Test Data Release: July 5, 2025
⏱️ Run Submission Deadline: July 20, 2025
🏆 Results Declaration: July 31, 2025
📝 Working Notes Due: August 31, 2025
📸 Camera-Ready Version: September 30, 2025
🔗 More Information & Registration: https://cmir-iitbhu.github.io/cmir/index.html
We warmly invite your participation in this timely and important task. Join us in tackling real-world challenges in multilingual, code-mixed information access and contribute to advancing research in socially grounded NLP.
Best regards,
The CMIR 2025 Team
Contact us: cmir....@gmail.com