Ruhr-Uni-Bochum
HGI

Copyright: HGI, stock.adobe.com: chinnarach

Bochum Team Best at Recognising Irony on Twitter

It is often difficult for algorithms to understand irony. However, this ability is often essential in order to judge whether, for example, a comment in social media is meant maliciously or not.

Photo Yu and Kolossa

Copyright: RUB, Marquard

A RUB team from Cognitive Signal Processing has won the PAN competition 2022. The task of the group led by Wentao Yu was to develop an algorithm that can judge whether people are expressing themselves ironically on Twitter and spreading stereotypes or not. Understanding irony is currently still difficult for machines. However, this ability is essential to be able to judge statements made by users in social media. Of 64 participants, the RUB team solved the task best: Wentao Yu, Benedikt Bönninghoff and Prof. Dr. Dorothea Kolossa were able to identify those accounts in the test data set that used irony to spread stereotypes against women or the LGBTQIA community with an accuracy of 99.44 percent.

PAN workshops
PAN has become one of the most important events for the field of authorship analysis. Since 2007, different competitions in the field of digital text forensics have been held, where researchers as well as practitioners around the world compare their methods for automatic analysis.

General note: In case of using gender-assigning attributes we include all those who consider themselves in this gender regardless of their own biological sex.