BRIDO: Bringing Democratic Order to Abstractive Summarization
This work addresses hallucination in LLMs for summarization, which is a critical issue for practical applications, though it is incremental as it builds on existing exposure bias mitigation methods.
The paper tackles hallucination in abstractive summarization by mitigating exposure bias through contrastive learning, achieving 6.25% and 3.82% improvements in consistency scores on XSum and CNN/DM datasets over the baseline BRIO.
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.