RST Discourse Treebank: 385 Annotated Wall Street Journal Articles

The RST Discourse Treebank was developed by researchers at the Information Sciences Institute (University of Southern California), the US Department of Defense, and the Linguistic Data Consortium. It consists of 385 Wall Street Journal articles from the Penn Treebank, annotated with discourse structure in the Rhetorical Structure Theory framework. The data is divided into a training set of 347 documents and a test set of 38 documents.

Use Cases

• Training discourse parsers based on annotated rhetorical relations and nuclearity.
• Developing text summarization models based on human-generated abstracts and extracts.
• Analyzing the structure of journalistic text based on annotated elementary discourse units and spans.
• Studying the application of Rhetorical Structure Theory to real-world documents.

Strengths

• 385 documents provide a foundational corpus for discourse analysis.
• Includes human-generated abstracts and extracts for 30 and 180 documents respectively, offering ground truth for summarization tasks.
• Annotations cover four key RST aspects: elementary discourse units, spans, nuclearity, and rhetorical relations.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Information Sciences Institute (University of Southern California), US Department of Defense, Linguistic Data Consortium.

Collection Method

Wall Street Journal articles from the Penn Treebank were annotated with discourse structure using a dedicated annotation tool.

Time Range

null

Freshness

null

Geography

null