Flickr30K Image Caption Quintets for Denotational Similarity

Aggregating Flickr30K image caption quintets used to compute denotational similarities for semantic inference tasks. It was created by embedding-data and last updated in August 2022.

Use Cases

• Construct a denotation graph over linguistic constituents and their visual denotations for semantic inference.
• Define novel denotational similarity metrics using the visual denotations of linguistic expressions.
• Compare the benefits of denotational similarities against distributional similarities for tasks requiring semantic inference.

Strengths

• Dataset is associated with the Flickr30K collection, a known benchmark in vision-language research.
• Tags indicate the dataset is in English and originates from the US region.
• The dataset is structured for a specific research purpose: computing denotational similarities.

Limitations

• The exact number of rows, columns, and sample data are unknown, limiting reproducibility.
• The dataset's size is unspecified, making it difficult to assess its scale for model training.
• The raw description is incomplete, requiring users to visit an external page for full details.

Provenance

Source

huggingface

Collection Method

Derived from the Flickr30K dataset for research on denotational similarity metrics.

Freshness

Last updated on 2022-08-02.

Geography

Regionus tag suggests a US focus, but specific coverage is unknown.