AI 中文总结
本文研究小型本地模型在信息抽取中精度与能耗的权衡，发现批处理与FP8量化显著节能，且文档类型决定最优输入表示：版面丰富用视觉-语言模型，近纯文本用纯文本模型配廉价解析器。
AI 中文摘要
信息抽取流水线应处理页面图像还是解析文本，这取决于文档类型，答案在版面谱系的两端截然不同。我们在排除（封闭）云服务的约束下研究这一权衡：面向隐私敏感文档，在本地部署小型（≤80亿参数）纯文本模型和视觉-语言模型，并在涵盖输入表示、模型家族和推理配置的设计空间中，同时评估精度和能耗。在近纯文本的Kleister-NDA合同和版面丰富的VRDU表单上进行基准测试，我们发现批处理是主导的能耗杠杆，在不损失精度的情况下将每页能耗降低38-85%，而FP8量化在逐个请求服务时节省27-32%，但一旦应用批处理，每页节省不到1毫瓦时（9-19%）。预处理主导剩余能耗：神经OCR每页能耗是经典OCR的17倍，且从未达到帕累托前沿。哪种表示胜出随文档类型而翻转：视觉-语言模型在版面丰富的文档上胜出，而小型纯文本模型配合廉价解析器在近纯文本上胜出，后者在精度和成本上均优于任何视觉-语言配置。我们的工作为节能、合规的本地信息抽取提供了具体指南。
英文摘要
Whether an information extraction pipeline should process page images or parsed text depends on the document, and the answer flips across the layout spectrum. We study this trade-off under a constraint that rules out (closed) cloud services: privacy-sensitive documents processed on-premise by small ($\le 8\mathrm{B}$ parameter) text-only and vision--language models, evaluated on both accuracy and energy over a design space spanning input representation, model family, and inference configuration. Benchmarking on the near-plain-text Kleister-NDA contracts and the layout-rich VRDU forms, we find that batching is the dominant energy lever, cutting energy per page by 38-85% at no cost in accuracy, while FP8 quantization saves 27-32% when requests are served one at a time but less than 1mWh per page (9-19%) once batching is applied. Preprocessing dominates what remains: neural OCR costs $17\times$ more energy per page than classical OCR and never reaches the Pareto frontier. Which representation wins flips with the type of document: vision--language models on layout-rich documents and small text-only models with a cheap parser on near-plain text, where they are both more accurate and cheaper than any vision--language configuration. Our work yields concrete guidelines for energy-efficient, privacy-compliant local information extraction.
CommentsAccepted to DocInsights at EMNLP 2026
