Multilingual and Multimodal LLMs for Everyday Knowledge

Birds-of-a-Feather Session at EMNLP 2025

This BoF brings together researchers and practitioners working on Multilingual and Multimodal LLMs for Everyday Knowledge with a focus on everyday and culturally grounded knowledge. As LLMs increasingly move beyond text into multimodal and spoken interactions, there is a growing need to understand how these models acquire, represent, and reason about culture, context, and everyday human knowledge, especially for low-resource languages and communities.

The session aims to foster open discussion around datasets, benchmarks, modeling strategies, and evaluation methodologies for culturally grounded multimodal AI.


📅 Event Details

  • Date: 06 November 2025
  • Time: 16:30 – 18:00
  • Room: A209

Program

🗓️ Program

Time Title Speaker / Presenter Mode
16:30 – 16:40 Opening
Slides
Firoj Alam, Qatar Computing Research Institute In-person
16:40 – 17:00 A Multilingual and Multimodal Framework for Culturally Grounded Spoken Visual QA
Slides
Shammur Absar Chowdhury, Qatar Computing Research Institute In-person
17:00 – 17:20 Grounding Multilingual Multimodal LLMs With Cultural Knowledge
Slides
Jean de Dieu Nyandwi, Carnegie Mellon University In-person
17:20 – 17:40 Pearl: A Multimodal Culturally-Aware Arabic Instruction Dataset
Slides
Muhammad Abdul-Mageed, University of British Columbia In-person
17:40 – 17:55 Open Discussion In-person
17:55 – 18:00 Closing In-person