1,500,000 images representing Wikipedia entities curated for the Visual Question Answering over Entities (ViQuAE) benchmark. These images serve as a visual knowledge base for tasks requiring models to link visual inputs to external structured information and natural language questions.
Use Cases
- Train entity-linking models to map visual features to specific Wikipedia entity identifiers
- Benchmark multimodal retrieval systems using the 1.5 million images as a search corpus for text-based queries
- Develop knowledge-based visual question answering models that require retrieving evidence from a visual corpus
Strengths
- 1,500,000 images mapped to distinct Wikipedia entities
- Supports the ViQuAE benchmark which includes 3,700 natural language question-answer pairs
- Released under the CC-BY-4.0 open-source license for research and commercial use