On the “easy” task of evaluating Chinese irony detection

Abstract

In this paper, we present a discussion on the problem in the evaluation of irony detection in Mandarin Chinese, especially due to the difficulties of finding an exhaustive definition and to the current lack of a gold standard for computational models. We describe some preliminary results of our experiments on an irony detection system for Chinese, and analyze examples of irony or other related phenomena that turned out to be challenging for NLP classifiers.