One of the challenges of developing a summarization model arises from the difficulty in measuring the factual inconsistency of the generated text. In this study, we reinterpret the decoder overconfidence-regularizing objective suggested in (Miao et al., 2021) as a hallucination risk measurement to better estimate the quality of generated summaries. We propose a reference-free metric, HaRiM+, which only requires an off-the-shelf summarization model to compute the hallucination risk based on token likelihoods. Deploying it requires no additional training of models or ad-hoc modules, which usually need alignment to human judgments. For summary-quality estimation, HaRiM+ records state-of-the-art correlation to human judgment on three summary-quality annotation sets: FRANK, QAGS, and SummEval. We hope that our work, which merits the use of summarization models, facilitates the progress of both automated evaluation and generation of summary.
@article{arxiv.2211.12118,
title = {HaRiM$^+$: Evaluating Summary Quality with Hallucination Risk},
author = {Seonil Son and Junsoo Park and Jeong-in Hwang and Junghwa Lee and Hyungjong Noh and Yeonsoo Lee},
journal= {arXiv preprint arXiv:2211.12118},
year = {2022}
}