Let's Stop Incorrect Comparisons in End-to-end Relation Extraction! - full paper at EMNLP 2020
by Bruno Taillé, Vincent Guigue, Geoffrey Scoutheeten, Patrick Gallinari
Despite efforts to distinguish three different evaluation setups, numerous end-to-end Relation Extraction articles present unreliable performance comparison to previous work. In this work, we identify several patterns of invalid comparisons in published papers and propose a small empirical study to quantify the impact of the most common mistake.
Collaboration with BNP Paribas