arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2408.09777cs.CL
收藏

使用多步流水线总结长篇监管文件

Summarizing long regulatory documents with a multi-step pipeline

Utrecht University(乌得勒支大学);Power2X(Power2X公司)

Mika Sie, Ruby Beek, Michiel Bots, Sjaak Brinkkemper, Albert Gatt

AI总结:

本文提出一种多步抽取-生成式架构用于总结长篇监管文件,发现两步架构的有效性取决于模型类型与上下文长度,并指出人工与自动评估存在差异,强调需根据模型特性选择总结策略。

摘要:

由于篇幅长且复杂,长篇监管文本的总结具有挑战性。为解决此问题,提出了一种多步抽取-生成式架构,以更有效地处理冗长的监管文件。在本文中,我们展示了用于总结长篇监管文本的两步架构的有效性因所使用的模型而异。具体而言,两步架构提升了 decoder-only 模型的性能。对于上下文长度较短的生成式 encoder-decoder 模型,抽取步骤的有效性各不相同;而对于长上下文 encoder-decoder 模型,抽取步骤会使其性能下降。本研究还强调了评估生成文本的挑战,人工评估和自动评估的不同结果证明了这一点。最值得注意的是,人工评估更青睐在法律文本上预训练的语言模型,而自动指标则将通用语言模型排名更高。结果强调,基于模型架构和上下文长度选择合适的总结策略的重要性。

英文摘要:

Due to their length and complexity, long regulatory texts are challenging to summarize. To address this, a multi-step extractive-abstractive architecture is proposed to handle lengthy regulatory documents more effectively. In this paper, we show that the effectiveness of a two-step architecture for summarizing long regulatory texts varies significantly depending on the model used. Specifically, the two-step architecture improves the performance of decoder-only models. For abstractive encoder-decoder models with short context lengths, the effectiveness of an extractive step varies, whereas for long-context encoder-decoder models, the extractive step worsens their performance. This research also highlights the challenges of evaluating generated texts, as evidenced by the differing results from human and automated evaluations. Most notably, human evaluations favoured language models pretrained on legal text, while automated metrics rank general-purpose language models higher. The results underscore the importance of selecting the appropriate summarization strategy based on model architecture and context length.

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。

↑