DeFacto is a dataset containing human demonstrations and feedback for improving factual consistency of text summarization.

The dataset is constructed with the following steps:

1. Detect errors: The annotator is required to evaluate a summary given the source document and decide if the summary is factually consistent.
2. Categorize errors: If the annotator decides the summary is not factually consistent, they are required to categorize the factual errors in the summary as either intrinsic or extrinsic.
3. Give explanation: The annotator is required to provide a natural language explanation on why the summary is factually consistent or not.
4. Provide evidence: The annotator is required to select a sentence from the source document as evidence to support their claims described in 3.
5. Write corrective instruction: The annotator is required to provide instructions of how to correct the original summary if they think it is not factually consistent. To enforce uniformity and reduce the noise in the instructions, we provide six templates for the annotators corresponding to different operations: Remove, Add, Replace, Modify, Rewrite, and Others. The annotators need to fill in the templates to generate the instructions.
6. Correct summary: Following the instruction in 5., the annotator is required to edit the initial summary to make it factually consistent with minimal, necessary modifications.

We use XSum as the target dataset and Pegasus as the pre-trained summarization model to generate the initial system outputs.

The dataset statistics are summarized below.

We provide the data files in ./data and a simple data loader data_loader.py.

Each line of the data files contain a data example stored in the Json format, with the following strucure:

 {
  "article": "input article",
  "abstract": "abstract/reference summary",
  "candidate": "candidate/initial system output",
  "doc_id": int,
  "has_error": true/false
  "intrinsic_error": true/false,
  "extrinsic_error": true/false,
  "feedback": {
    "summary": "human-corrected summary",
    "evidence": "selected sentence from the input article",
    "explanation": "natural language explanation",
    "instruction": "concatenated instructions",
    "instruction_list": ["list of instructions", ],
  },
}

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact [email protected] with any additional questions or comments.

This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party's policies.