WSL: Word Sense Linking Dataset with WordNet Annotations

A dataset for Word Sense Linking (WSL), requiring systems to disambiguate text spans to senses from a reference inventory. Annotations are provided as sense keys from WordNet, a large lexical database of English, and all terms have been fully manually annotated. The dataset sentences are taken from the ALL split of the Raganato framework and was authored by Babelscape.

Use Cases

• Training word sense disambiguation models based on manually annotated sense keys.
• Benchmarking semantic linking systems against a reference inventory from WordNet.
• Evaluating the performance of language models on lexical ambiguity resolution tasks.

Strengths

• All terms in the dataset have been fully manually annotated.
• Annotations use sense keys from WordNet, a large, established lexical database.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Babelscape

Collection Method

Sentences are taken from the ALL split of the Raganato framework.

Freshness

Last updated 2024-12-13 08:44:00; freshness should be verified.