Value Retrieval with Arbitrary Queries for Form-like Documents
This work addresses the problem of automating form processing for users who handle documents, offering a more flexible approach than previous methods, though it appears incremental in advancing document understanding techniques.
The paper tackles the problem of retrieving values from form-like documents using arbitrary queries, rather than a fixed set of fields, to reduce human effort in form processing. It introduces a method that predicts target values based on layout and semantic understanding, and shows significant performance improvements, with a 17% F1 score boost over state-of-the-art pre-training methods using a proposed SimpleDLM strategy.
We propose value retrieval with arbitrary queries for form-like documents to reduce human effort of processing forms. Unlike previous methods that only address a fixed set of field items, our method predicts target value for an arbitrary query based on the understanding of the layout and semantics of a form. To further boost model performance, we propose a simple document language modeling (SimpleDLM) strategy to improve document understanding on large-scale model pre-training. Experimental results show that our method outperforms previous designs significantly and the SimpleDLM further improves our performance on value retrieval by around 17% F1 score compared with the state-of-the-art pre-training method. Code is available at https://github.com/salesforce/QVR-SimpleDLM.