LT group is winner of GermEval 2018Offensive language detection with neural networks
21 September 2018, by LT group
Together with a team from TU Wien, LT group won the GermEval 2018 shared task on offensive language detection in Twitter
This year's shared task of the GermEval workshop on offensive language detection in Twitter data had two winners. Task 2, a fine-grained classification of tweets into four categories (abuse, insult, profanity, other) was won by the LT group team with a BiLSTM-CNN neural network architecture in a transfer learning setup. Besides LT, ten other groups participated in the shared task.
Task 1, a coarse-grained automatic classification of tweets into offensive or non-offensive was won by TU Wien. LT achieved place four out of 20 participating teams.
* More information and all results of GermEval 2018
* Proceedings of the GermEval 2018 Workshop
This effort is part of our Forum 4.0 project as pasrt of the ahoi.digital platform.