Table VQA: Bilingual Visual Question Answering on Scientific Table Images

The dataset integrates table images from the AFTdb (Arxiv Figure Table Database) curated by cmarkea. Each image is paired with LaTeX source code and linked to an average of ten questions and answers, half in English and half in French. Questions and answers were generated using Gemini 1.5 Pro and Claude 3.5 Sonnet, and the dataset was last updated on 2024-09-26.

Use Cases

• Train visual question answering models based on images of scientific tables.
• Develop models for table structure recognition based on the paired LaTeX source code.
• Benchmark multilingual VQA performance based on the bilingual English and French Q&A pairs.
• Fine-tune large language models for document understanding tasks based on the table image and text pairs.

Strengths

• Pairs each table image with corresponding LaTeX source code.
• Each image is linked to an average of ten question and answer pairs.
• Q&A pairs are evenly split between English and French languages.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Dataset size, file formats, and license information are unknown.

Provenance

Source

AFTdb (Arxiv Figure Table Database) curated by cmarkea.

Collection Method

Q&A pairs were generated using Gemini 1.5 Pro and Claude 3.5 Sonnet.

Freshness

Last updated 2024-09-26 11:53:12; freshness should be verified.