Authors: Chang-Uk Jeong

Description: The recognized texts will be aligned by analyzing spacial correlation, then transformed into a training dataset for fine-tuning the BERT-Base model on MRC.