BRIDO: Bringing Democratic Order to Abstractive Summarization

TL;DR

This paper introduces BRIDO, a novel approach to reduce hallucinations in abstractive summarization by leveraging contrastive learning and exposure bias mitigation, leading to more consistent summaries.

Contribution

It proposes a new model that aligns exposure bias mitigation with hallucination reduction using contrastive learning, improving summary consistency.

Findings

Achieved 6.25% improvement on XSum dataset.

Achieved 3.82% improvement on CNN/DM dataset.

Reduced hallucination in abstractive summarization.

Abstract

Hallucination refers to the inaccurate, irrelevant, and inconsistent text generated from large language models (LLMs). While the LLMs have shown great promise in a variety of tasks, the issue of hallucination still remains a major challenge for many practical uses. In this paper, we tackle the issue of hallucination in abstract text summarization by mitigating exposure bias. Existing models targeted for exposure bias mitigation, namely BRIO, aim for better summarization quality in the ROUGE score. We propose a model that uses a similar exposure bias mitigation strategy but with a goal that is aligned with less hallucination. We conjecture that among a group of candidate outputs, ones with hallucinations will comprise the minority of the whole group. That is, candidates with less similarity with others will have a higher chance of containing hallucinated content. Our method uses this aspect and utilizes contrastive learning, incentivizing candidates with high inter-candidate ROUGE scores. We performed experiments on the XSum and CNN/DM summarization datasets, and our method showed 6.25% and 3.82% improvement, respectively, on the consistency G-Eval score over BRIO.