VULCA-Bench: Bilingual Multicultural Art-Critique Corpus with 7,236 Samples

VULCA-Bench is a bilingual multicultural art-critique corpus containing 7,236 multimodal samples, with 7,234 including embedded images. It covers eight cultural traditions and uses a schema with 236 cultural dimensions. The dataset was created by author harryHURRY and last updated on April 30, -2026.

Use Cases

• Training multimodal AI models for art critique based on bilingual text and image pairs.
• Benchmarking cultural bias in vision-language models based on the eight specified cultural traditions.
• Analyzing cross-cultural aesthetic judgments based on the 236-dimension labelling schema.
• Fine-tuning models for image-text alignment tasks based on the dataset's multimodal configuration.

Strengths

• Contains 7,236 total samples, providing a substantial corpus for model training.
• Covers eight distinct cultural traditions (Chinese, Western, Japanese, Korean, Islamic, Indian, Mural, Hermitage).
• Employs a detailed labelling schema with 236 cultural dimensions.
• Most samples (7,234) are multimodal with embedded images.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Two mural records are metadata-only with orphan image paths, indicating potential data integrity issues.
• The description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

huggingface

Collection Method

Likely collected and annotated for art critique and cultural analysis.

Time Range

The dataset page was last updated in -2026, but the temporal coverage of the content is unknown.

Freshness

Last updated 2026-04-30 21:31:19; freshness should be verified.

Geography

Global, covering eight specified cultural traditions.