在不牺牲摘要质量的前提下提升生成式摘要的事实性
Improving Factuality of Abstractive Summarization without Sacrificing Summary Quality
AI总结：
提出EFACTSUM候选摘要生成与排序技术，结合对比学习框架与排序策略，在不牺牲摘要质量的前提下显著提升生成式摘要的事实性。
摘要：
提升生成式摘要的事实一致性一直是一个被广泛研究的课题。然而，先前大多数关于训练事实性感知模型的工作都忽略了其对摘要质量的负面影响。我们提出了EFACTSUM（即Effective Factual Summarization），这是一种候选摘要生成与排序技术，旨在不牺牲摘要质量的前提下提升摘要的事实性。我们表明，将对比学习框架与我们改进的候选摘要结合使用，能够在事实性和基于相似度的指标上带来显著提升。具体而言，我们提出了一种排序策略，有效地结合了两个指标，从而防止训练期间出现任何冲突。使用我们方法训练的模型在XSUM数据集上的FactCC指标上比基础模型表现出高达6个绝对点的提升，在CNN/DM上表现出11个绝对点的提升，且未对基于相似度的指标或摘要的抽象性产生负面影响。
英文摘要：
Improving factual consistency of abstractive summarization has been a widely studied topic. However, most of the prior works on training factuality-aware models have ignored the negative effect it has on summary quality. We propose EFACTSUM (i.e., Effective Factual Summarization), a candidate summary generation and ranking technique to improve summary factuality without sacrificing summary quality. We show that using a contrastive learning framework with our refined candidate summaries leads to significant gains on both factuality and similarity-based metrics. Specifically, we propose a ranking strategy in which we effectively combine two metrics, thereby preventing any conflict during training. Models trained using our approach show up to 6 points of absolute improvement over the base model with respect to FactCC on XSUM and 11 points on CNN/DM, without negatively affecting either similarity-based metrics or absractiveness.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
