Can Cognitive Scientists Help Computers Recognize Irony?

This workshop aims to bring together researchers in cognitive science and computer science with a shared interest in irony. Irony detection is an especially difficult problem for natural language processing. Unlike other types of classification tasks, the difficulty of identifying irony is not alleviated by ‘throwing more data’ at the problem. Rather, it seems a different kind of data is needed: contextual data. We also need new models that can exploit this data. In this workshop, which capitalizes on the collocation of the AAAI and CogSci conferences, we invite cognitive scientists and computer scientists to engage in a dialogue around new machine learning models for irony detection and new methods and tools for testing predictions of cognitive theories of irony against large-scale data sets.