Authors: Hanqin Wang, Jie Qin, Fan Zhu, Li Liu, and Ling Shao (Inception Institute of Artificial Intelligence)
Description: In this method, we extract the key information from the text from the following two perspectives. In terms of the 'company', 'date', and 'address' information, we fine-tune the BERT model on the training set to obtain several one-vs-all classifiers. As for the 'total' information, we take advantage of the text localization results. Specifically, we extract the target numbers close to the specific text indicating the corresponding information.