An unofficial code reproduction in the field of event extraction of EMNLP-19 paper "Event Detection with Multi-Order Graph Convolution and Aggregated Attention"
Hi there, very excellent reproduction work, and thank you for your sharing!
I got that there is a little inconsistency with the original MOGANED paper in labeling schema. And I wonder whether the change of labeling schema (from BIO labeling to non-BIO) would loose the evaluation criteria, and then resulting in an increase of Precision, Recall, F1? If so, is it not comparable with other people's papers due to the looser criteria?
Hi there, very excellent reproduction work, and thank you for your sharing!
I got that there is a little inconsistency with the original MOGANED paper in labeling schema. And I wonder whether the change of labeling schema (from BIO labeling to non-BIO) would loose the evaluation criteria, and then resulting in an increase of Precision, Recall, F1? If so, is it not comparable with other people's papers due to the looser criteria?
Hope I will receive your reply soon, thank you!