Keeping Their Words: Direct and Indirect Chinese Quote Attribution from Newspapers

Kuan-Lin Lee, Yu-Chung Cheng, Pai-Lin Chen, Hen-Hsen Huang

Abstract: Quote attribution plays an important role in emerging research topics such as fact checking, stance detection, and argument mining. This work explores Chinese quote attribution from newspapers. Both direct and indirect quotes are addressed by a text-encoder based sequence labeling model. We create a dataset for empirical analysis. Experimental results show the effectiveness of our model.