FormNet: Structural Encoding beyond Sequential Modeling in Form Document Information Extraction
This addresses the challenge of extracting information from form documents with varied layouts, offering a more efficient solution for document understanding tasks.
The paper tackled the problem of suboptimal token serialization in form-like documents by proposing FormNet, a structure-aware sequence model that leverages spatial relationships and graph convolutions, achieving new state-of-the-art performance on CORD, FUNSD, and Payment benchmarks.
Sequence modeling has demonstrated state-of-the-art performance on natural language and document understanding tasks. However, it is challenging to correctly serialize tokens in form-like documents in practice due to their variety of layout patterns. We propose FormNet, a structure-aware sequence model to mitigate the suboptimal serialization of forms. First, we design Rich Attention that leverages the spatial relationship between tokens in a form for more precise attention score calculation. Second, we construct Super-Tokens for each word by embedding representations from their neighboring tokens through graph convolutions. FormNet therefore explicitly recovers local syntactic information that may have been lost during serialization. In experiments, FormNet outperforms existing methods with a more compact model size and less pre-training data, establishing new state-of-the-art performance on CORD, FUNSD and Payment benchmarks.