AlignSum：面向人类摘要偏好对齐的数据金字塔分层微调
AlignSum: Data Pyramid Hierarchical Fine-tuning for Aligning with Human Summarization Preference
AI总结：
本文提出AlignSum框架，通过构建数据金字塔、高斯重采样和两阶段分层微调，使BART-Large等预训练语言模型在自动与人类评估中超越175B GPT-3，显著提升摘要生成与人类偏好的对齐。
摘要：
文本摘要任务通常使用预训练语言模型（PLMs）来拟合各种标准数据集。尽管这些PLMs在自动评估中表现优异，但它们在人类评估中常常表现不佳，表明其生成的摘要与人类摘要偏好之间存在偏差。这种差异可能源于微调数据集质量较低，以及能够反映真实人类偏好的高质量人工标注数据有限。为了解决这一挑战，我们提出了一种新颖的人类摘要偏好对齐框架AlignSum。该框架由三部分组成：首先，我们构建了一个包含抽取式、生成式和人工标注摘要数据的数据金字塔（Data Pyramid）。其次，我们执行高斯重采样（Gaussian Resampling）以去除长度极端的摘要。最后，我们在高斯重采样之后，利用数据金字塔实施两阶段分层微调。我们将AlignSum应用于人工标注的CNN/DailyMail和BBC XSum数据集上的PLMs。实验表明，借助AlignSum，BART-Large等PLMs在自动评估和人类评估中均超越了175B参数的GPT-3。这表明AlignSum显著增强了语言模型与人类摘要偏好的对齐程度。
英文摘要：
Text summarization tasks commonly employ Pre-trained Language Models (PLMs) to fit diverse standard datasets. While these PLMs excel in automatic evaluations, they frequently underperform in human evaluations, indicating a deviation between their generated summaries and human summarization preferences. This discrepancy is likely due to the low quality of fine-tuning datasets and the limited availability of high-quality human-annotated data that reflect true human preference. To address this challenge, we introduce a novel human summarization preference alignment framework AlignSum. This framework consists of three parts: Firstly, we construct a Data Pymarid with extractive, abstractive, and human-annotated summary data. Secondly, we conduct the Gaussian Resampling to remove summaries with extreme lengths. Finally, we implement the two-stage hierarchical fine-tuning with Data Pymarid after Gaussian Resampling. We apply AlignSum to PLMs on the human-annotated CNN/DailyMail and BBC XSum datasets. Experiments show that with AlignSum, PLMs like BART-Large surpass 175B GPT-3 in both automatic and human evaluations. This demonstrates that AlignSum significantly enhances the alignment of language models with human summarization preferences.
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。
