论文标题
视觉丰富的文档的增强策略
An Augmentation Strategy for Visually Rich Documents
论文作者
论文摘要
许多业务工作流程需要从类似表单的文档（例如银行对帐单，提单，采购订单等）中提取重要字段。仅在接受大型数据集培训时，最新的自动化此任务的技术才能很好地工作。在这项工作中，我们提出了一种新颖的数据增强技术，以提高培训数据稀缺时的性能，例如10-250个文件。我们称我们称为fieldswap的技术是通过将源字段的关键短语与目标字段的关键短语交换来生成目标字段的新合成示例，以用于培训。我们证明这种方法可以在提取性能方面产生1-7 F1点的改善。
Many business workflows require extracting important fields from form-like documents (e.g. bank statements, bills of lading, purchase orders, etc.). Recent techniques for automating this task work well only when trained with large datasets. In this work we propose a novel data augmentation technique to improve performance when training data is scarce, e.g. 10-250 documents. Our technique, which we call FieldSwap, works by swapping out the key phrases of a source field with the key phrases of a target field to generate new synthetic examples of the target field for use in training. We demonstrate that this approach can yield 1-7 F1 point improvements in extraction performance.
