DocumentNet: Bridging the Data Gap in Document Pre-Training
This addresses the data gap for document understanding tasks in enterprise AI, though it is incremental as it builds on existing pre-training methods by providing a new dataset.
The paper tackles the scarcity of publicly available data for Visually-rich Document Entity Retrieval (VDER) by proposing DocumentNet, a method to collect massive-scale, weakly labeled data from the web, resulting in significant improvements on VDER tasks, including in few-shot learning settings.
Document understanding tasks, in particular, Visually-rich Document Entity Retrieval (VDER), have gained significant attention in recent years thanks to their broad applications in enterprise AI. However, publicly available data have been scarce for these tasks due to strict privacy constraints and high annotation costs. To make things worse, the non-overlapping entity spaces from different datasets hinder the knowledge transfer between document types. In this paper, we propose a method to collect massive-scale and weakly labeled data from the web to benefit the training of VDER models. The collected dataset, named DocumentNet, does not depend on specific document types or entity sets, making it universally applicable to all VDER tasks. The current DocumentNet consists of 30M documents spanning nearly 400 document types organized in a four-level ontology. Experiments on a set of broadly adopted VDER tasks show significant improvements when DocumentNet is incorporated into the pre-training for both classic and few-shot learning settings. With the recent emergence of large language models (LLMs), DocumentNet provides a large data source to extend their multi-modal capabilities for VDER.