This study investigates the automatic detection and classification of persuasion techniques across five Slavic languages (Bulgarian, Croatian, Polish, Russian, and Slovenian), addressing two subtasks: binary detection of persuasion techniques in text fragments (Subtask 1) and multi-label classification of specific technique types (Subtask 2). To overcome limited training resources, we implemented a multi-level cross-lingual augmentation strategy utilizing GPT-4o for non-Slavic to Slavic conversion and intra-Slavic language migration. We employ XLM-RoBERTa architecture with two LLM-enhanced variants that use explanations to improve classification performance. The experimental results demonstrate varied performance across languages and tasks, with our approach achieving first place in the Russian subtask 1 and second place in Bulgarian subtask 2, confirming that larger parameter models excel in complex classification tasks. These findings highlight the significant potential of LLMs for enhancing multilingual classification and the persistent difficulties in ensuring consistent cross-linguistic performance.

Le suivi efficace de l’acquisition des compétences dans les programmes de l’alternance, présente des défis importants pour la technologie éducative. Cet article présente un nouvel agent conversationnel intégré dans un livret de formation numérique qui relève ces défis grâce à une architecture multimodale. Notre système intègre (1) un graphe de connaissances spécifique à un domaine, lié à des référentiels de compétences, (2) des grands modèles de langage (LLM) et (3) un composant génératif basé sur BERT. Cette approche hybride permet à la fois une représentation structurée des trajectoires d’apprentissage et des capacités d’interaction en langage naturel, ce qui permet un suivi nuancé des progrès et des interventions personnalisées. L’évaluation empirique démontre que le système fournit un retour d’information contextuellement pertinent qui s’adapte aux modèles d’apprentissage individuels, ce qui permet une acquisition plus efficace des compétences.

We present Team INSALyon2’s agentic approach to SemEval-2025 Task 10 Subtask 2, which focuses on the multi-label classification of narratives in news articles across five languages. Our system employs a zero-shot architecture where specialized Large Language Model (LLM) agents handle binary classification tasks for individual narrative/subnarrative labels, with a meta-agent aggregating these decisions into final multi-label predictions. Instead of fine-tuning on the dataset, we leverage AutoGen to orchestrate multiple GPT-based agents, each responsible for detecting specific narrative/subnarrative types in a modular framework. This agent-based approach naturally handles the challenge of multi-label classification by enabling parallel decisions across the two-level taxonomy. Experiments on the English subset demonstrate strong performance with our system achieving F1_macro_coarse = 0.513, F1_sample = 0.406, securing third place in the competition. Our findings show that zero-shot agentic approaches can be competitive in complex classification tasks.

In this paper, we propose a BERT-based hierarchical text classification framework to address the challenges of training multi-level classification tasks. As part of the SemEval-2025 Task 10 challenge (Subtask 2), the framework performs fine-grained text classification by training dedicated sub-category classifiers for each top-level category. Experimental results demonstrate the feasibility of the proposed approach in multi-class text classification tasks.

Medical discourse within the professional community has undeservingly received very sparse researchers’ attention. Medical professional discourse exists offline and online. We carried out sentiment analysis on titles and text descriptions of materials published on the Russian portal Mir Vracha (90,000 word forms approximately). The texts were generated by and for physicians. The materials include personal narratives describing participants’ professional experience, participants’ opinions about pandemic news and events in the professional sphere, and Russian reviews and discussion of papers published in international journals in English. We present the first results and discussion of the sentiment analysis of Russian online medical discourse. Based on the results of sentiment analysis and discourse analysis, we described the emotions expressed in the forum and the linguistic means the forum participants used to verbalise their attitudes and emotions while discussing the Covid-19 pandemic. The results showed prevalence of neutral texts in the publications since the medical professionals are interested in research materials and outcomes. In the discussions and personal narratives, the forum participants expressed negative sentiments by colloquial words and figurative language.

Social media have become a valuable source of information. However, its power to shape public opinion can be dangerous, especially in the case of misinformation. The existing studies on misinformation detection hypothesise that the initial message is fake. In contrast, we focus on information distortion occurring in cascades as the initial message is quoted or receives a reply. We show a significant topic shift in information cascades on Twitter during the Covid-19 pandemic providing valuable insights for the automatic analysis of information distortion.