arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2305.04460cs.CLcs.AI
收藏

语言无关的神经符号语义解析用于表单理解

Language Independent Neuro-Symbolic Semantic Parsing for Form Understanding

Monash University（莫纳什大学）

Bhanu Prakash Voutharoja, Lizhen Qu, Fatemeh Shiri

备注：
Accepted to ICDAR 2023
阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出语言无关的图神经网络LAGNN，仅利用布局信息解析表单为实体-关系图，结合ILP推理提升性能，实现跨语言表单理解。

摘要：

近期关于表单理解的研究大多采用多模态变换器或大规模预训练语言模型。这些模型需要大量数据进行预训练。相比之下，人类通常仅通过查看布局就能识别表单中的键值对，即使他们不理解所使用的语言。此前尚无研究探讨仅凭布局信息对表单理解的帮助程度。因此，我们提出了一种针对扫描表单的独特实体-关系图解析方法，称为LAGNN，一种语言无关的图神经网络模型。我们的模型将表单解析为词-关系图，以联合识别实体和关系，并降低推理的时间复杂度。随后，该图通过确定性规则转换为全连接的实体-关系图。我们的模型仅考虑布局信息中边界框之间的相对间距，以促进跨语言的轻松迁移。为了进一步提升LAGNN的性能，并实现实体-关系图与词-关系图之间的同构，我们采用了基于整数线性规划（ILP）的推理。代码公开于https://github.com/Bhanu068/LAGNN

英文摘要：

Recent works on form understanding mostly employ multimodal transformers or large-scale pre-trained language models. These models need ample data for pre-training. In contrast, humans can usually identify key-value pairings from a form only by looking at layouts, even if they don't comprehend the language used. No prior research has been conducted to investigate how helpful layout information alone is for form understanding. Hence, we propose a unique entity-relation graph parsing method for scanned forms called LAGNN, a language-independent Graph Neural Network model. Our model parses a form into a word-relation graph in order to identify entities and relations jointly and reduce the time complexity of inference. This graph is then transformed by deterministic rules into a fully connected entity-relation graph. Our model simply takes into account relative spacing between bounding boxes from layout information to facilitate easy transfer across languages. To further improve the performance of LAGNN, and achieve isomorphism between entity-relation graphs and word-relation graphs, we use integer linear programming (ILP) based inference. Code is publicly available at https://github.com/Bhanu068/LAGNN

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。

↑