Despite the seeming success of contemporary grounded text generation systems, they often tend to generate factually inconsistent text with respect to their input. This phenomenon is emphasized in tasks like summarization, in which the generated summaries should be corroborated by their source article. In this work, we leverage recent progress on textual entailment models to directly address this problem for abstractive summarization systems. We use reinforcement learning with reference-free, textual entailment rewards to optimize for factual consistency and explore the ensuing trade-offs, as improved consistency may come at the cost of less informative or more extractive summaries. Our results, according to both automatic metrics and human evaluation, show that our method considerably improves the faithfulness, salience, and conciseness of the generated summaries.
@article{arxiv.2306.00186,
title = {Factually Consistent Summarization via Reinforcement Learning with Textual Entailment Feedback},
author = {Paul Roit and Johan Ferret and Lior Shani and Roee Aharoni and Geoffrey Cideron and Robert Dadashi and Matthieu Geist and Sertan Girgin and Léonard Hussenot and Orgad Keller and Nikola Momchev and Sabela Ramos and Piotr Stanczyk and Nino Vieillard and Olivier Bachem and Gal Elidan and Avinatan Hassidim and Olivier Pietquin and Idan Szpektor},
journal= {arXiv preprint arXiv:2306.00186},
year = {2023}
}