CVQA: Culturally-diverse Multilingual Visual Question Answering Benchmark
Published in Proceedings of the 38th International Conference on Neural Information Processing Systems (NeurIPS 2024), 2024
Visual Question Answering (VQA) is a crucial task in multimodal AI, often used to test vision-language models on understanding visual and linguistic information. However, existing benchmarks predominantly focus on Western and English-centric scenarios. CVQA introduces a culturally-diverse multilingual benchmark covering dozens of languages and diverse geographic regions to evaluate model generalization and cultural grounding across the world.