Appeal for attention at SemEval-2023 Task 3: Data augmentation and extension strategies for detection of online news persuasion techniques

Abstract

In this paper, we proposed and explored the impact of four different dataset augmentation and extension strategies that we used for solving the subtask 3 of SemEval-2023 Task 3: multilabel persuasion techniques classification in a multi-lingual context. We consider two types of augmentation methods (one based on a modified version of synonym replacement and one based on translations) and two ways of extending the training dataset (using filtered data generated by GPT-3 and using a dataset from a previous competition). We studied the effects of the aforementioned techniques by using the augmented and/or extended training dataset to fine-tune a pretrained XLM-RoBERTa-Large model. Using the augmentation methods alone, we managed to obtain 3rd place for English, 13th place for Italian and between the 5th to 9th places for the other 7 languages during the competition.