Success of doctoral students in the competition to the international workshop SemEval

In the winter semester of 2022/2023, PhD students of the ICT&Design Doctoral School, Arkadiusz Modzelewski and Witold Sosnowski, won qualification for the 17th edition of the SemEval 2023 International Semantic Evaluation Workshop. They participated in a task on Category Detection, Framing and Persuasion Techniques in Web Messages in a Multilingual Configuration.

As part of the task, they investigated the ability of natural language processing (NLP) systems to detect genres and persuasion techniques in multiple languages, and conducted experiments with various data augmentation techniques, including machine translation and text generation.

To do this, they used Davinci's OpenAI GPT-3 language model to create synthetic texts for species detection and relied on augmenting the dataset by translating the texts using a translator DeepL as part of their detection of persuasion techniques.

The tuned models that used the expanded data ranked in the top 10 in all languages, indicating the effectiveness of the proposed approach. 

The species detection models showed excellent results, securing first, second and third place in Spanish, German and Italian, respectively.

In turn, one of the models for detecting persuasion techniques ranked third in Polish. Their contributions to science include a system architecture that uses DeepL and GPT-3 to extend the data to detect both species and persuasion techniques.

The proposed methods are described in the article: "DSHacker at SemEval-2023 Task 3: Genres and Persuasion Techniques Detection with Multilingual Data Augmentation through Machine Translation and Text Generation," which has been accepted for publication. The SemEval workshop at which the results of the competition will be presented will be held as part of the ACL-2023 conference in Toronto, Canada. 

ICT & Design Doctoral School has announced recruitment for the 2023/24 academic year. More information here.