• DOI:10.18653/v1/2020.emnlp-main.750
Corpus ID: 204976362

Evaluating the Factual Consistency of Abstractive Text Summarization

@article{Kryscinski2020EvaluatingTF,
  title={Evaluating the Factual Consistency of Abstractive Text Summarization},
  author={Wojciech Kryscinski and B. McCann and Caiming Xiong and R. Socher},
  journal={ArXiv},
  year={2020},
  volume={abs/1910.12840}
}

Wojciech Kryscinski, B. McCann, +1 author R. Socher

Published 2020

Computer Science

ArXiv

Currently used metrics for assessing summarization algorithms do not account for whether summaries are factually consistent with source documents. We propose a weakly-supervised, model-based approach for verifying factual consistency and identifying conflicts between source documents and a generated summary. Training data is generated by applying a series of rule-based transformations to the sentences of source documents. The factual consistency model is then trained jointly for three tasks: 1… CONTINUE READING

View on ACL

arxiv.org

Save to Library

Create Alert

Cite

Launch Research Feed