用掩码并行生成，让文档信息提取快36倍
Up to 36x Speedup: Mask-based Parallel Inference Paradigm for Key Information Extraction in MLLMs
- 用[mask]占位符替代逐个生成，实现多字段并行输出
- 实验显示速度提升5至36倍，准确率几乎不变
- 适合需要快速处理大量文档的工业场景
从视觉丰富文档（VrDs）中提取关键信息是重要任务，近年大语言模型（LLMs）和多模态大语言模型（MLLMs）展现出强大潜力。然而，其依赖自回归推理、逐词生成，导致效率瓶颈，尤其在需提取多个语义独立字段时更为明显。为此，我们提出PIP：面向关键信息提取的并行推理范式。通过使用[masked] token作为所有目标值的占位符，实现单次前向传播中同时生成全部字段。为支持该范式，我们设计了定制化的掩码预训练策略，并构建大规模监督数据集。实验表明，相比传统自回归基线模型，PIP模型实现5-36倍的推理加速，性能损失可忽略。PIP显著提升效率的同时保持高精度，为可扩展、实用的现实世界关键信息提取方案铺平道路。
原文摘要 · Abstract (English)
Key Information Extraction (KIE) from visually-rich documents (VrDs) is a critical task, for which recent Large Language Models (LLMs) and Multi-Modal Large Language Models (MLLMs) have demonstrated strong potential. However, their reliance on autoregressive inference, which generates outputs sequentially, creates a significant efficiency bottleneck, especially as KIE tasks often involve extracting multiple, semantically independent fields. To overcome this limitation, we introduce PIP: a Parallel Inference Paradigm for KIE. Our approach reformulates the problem by using "[mask]" tokens as placeholders for all target values, enabling their simultaneous generation in a single forward pass. To facilitate this paradigm, we develop a tailored mask pre-training strategy and construct large-scale supervised datasets. Experimental results show that our PIP-models achieve a 5-36x inference speedup with negligible performance degradation compared to traditional autoregressive base models. By substantially improving efficiency while maintaining high accuracy, PIP paves the way for scalable and practical real-world KIE solutions.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。