myFoodQA: A Multimodal Dataset for Evaluating Cultural and Visual Reasoning in Myanmar Gastronomy
Published in 7th Workshop on Innovation Initiatives in NLP/AI/Education (IINAE 2025), iSAI-NLP 2025, 2025
The myFoodQA dataset captures 1,500+ curated Myanmar dishes with paired visual and textual descriptions sourced across regions. We annotate multiple reasoning-intensive tasks (cultural context, ingredient-level grounding, recommendations, and visual QA) to evaluate multimodal systems in low-resource settings. The paper details our collection pipeline, evaluation splits, and baseline models, and it has been accepted as a FULL paper for presentation at the 7th Workshop on Innovation Initiatives in NLP/AI/Education (IINAE 2025) co-located with iSAI-NLP 2025.
Recommended citation: Pyae Linn et al. (2025). "myFoodQA: A Multimodal Dataset for Evaluating Cultural and Visual Reasoning in Myanmar Gastronomy." IINAE 2025 @ iSAI-NLP 2025. FULL paper #253.
Download Paper