Cofced: Evidence-Based Fake News Detection from COLING 2022

Cofced provides text data for explainable fake news detection using a coarse-to-fine evidence distillation approach. Developed by Nicozwy for the COLING 2022 conference, the dataset focuses on providing verifiable evidence for news claims.

Use Cases

• Developing explainable fake news classifiers using the evidence-distillation labels
• Testing evidence retrieval algorithms against the news claims provided in the text
• Implementing the cascaded neural network architecture described in the associated COLING 2022 paper

Strengths

• Peer-reviewed origin from the COLING 2022 conference
• Focuses on explainability through evidence-distillation labels
• Released under the permissive Apache-2.0 license

Limitations

• Specific record counts and column schemas are not explicitly documented in the repository metadata
• Potential for temporal staleness as news-related data is highly time-sensitive

Provenance

Source

COLING 2022: A Coarse-to-fine Cascaded Evidence-Distillation Neural Network for Explainable Fake News Detection