A5081530197- Misinformation and Its Impacts
- Reflective Practices in Education
- Natural Language Processing Techniques
- Intelligent Tutoring Systems and Adaptive Learning
- Hate Speech and Cyberbullying Detection
- Topic Modeling
- Authorship Attribution and Profiling
- Opinion Dynamics and Social Influence
- linguistics and terminology studies
- Generative Adversarial Networks and Image Synthesis
- Social Media and Politics
- Intelligence, Security, War Strategy
- Crime Patterns and Interventions
- Literature, Language, and Rhetoric Studies
- Digital Communication and Language
- Advanced Image and Video Retrieval Techniques
- Visual Attention and Saliency Detection
- Library Science and Information Systems
- Security, Politics, and Digital Transformation
Freie Universität Berlin
2022-2023
University of Bamberg
2023
Abstract The full-scale conflict between the Russian Federation and Ukraine generated an unprecedented amount of news articles social media data reflecting opposing ideologies narratives. These polarized campaigns have led to mutual accusations misinformation fake news, shaping atmosphere confusion mistrust for readers worldwide. This study analyses how affected mirrored public opinion during first month war using Telegram channels in Ukrainian, Russian, Romanian, French English. We propose...
We introduce VIBA, a novel approach for explainable video classification by adapting Information Bottlenecks Attribution (IBA) to sequences. While most traditional explainability methods are designed image models, our IBA framework addresses the need in temporal models used analysis. To demonstrate its effectiveness, we apply VIBA deepfake detection, testing it on two architectures: Xception model spatial features and VGG11-based capturing motion dynamics through optical flow. Using custom...
We discuss how desirable it is that Large Language Models (LLMs) be able to adapt or align their language behavior with users who may diverse in use. User diversity come about among others due i) age differences; ii) gender characteristics, and/or iii) multilingual experience, and associated differences processing consider potential consequences for usability, communication, LLM development.
In the Russo-Ukrainian war, propaganda is produced by Russian state-run news outlets for both international and domestic audiences. Its content form evolve change with time as war continues. This constitutes a challenge to moderation tools based on machine learning when data used training current start differ significantly. this follow-up study, we evaluate our previous BERT SVM models that classify Pro-Kremlin from Pro-Western stance, trained articles telegram posts at of 2022, new 2023...
Many European citizens become targets of the Kremlin propaganda campaigns, aiming to minimise public support for Ukraine, foster a climate mistrust and disunity, shape elections (Meister, 2022). To address this challenge, we developed ''Check News in 1 Click'', first NLP-empowered pro-Kremlin detection application available 7 languages, which provides lay user with feedback on their news, explains manipulative linguistic features keywords. We conducted study, analysed entries models'...
Written reflective practice is a regular exercise pre-service teachers perform during their higher education. Usually, lecturers are expected to provide individual feedback, which can be challenging task on basis. In this paper, we present the first open-source automated feedback tool based didactic theory and implemented as hybrid AI system. We describe components discuss advantages disadvantages of our system compared state-of-art generative large language models. The main objective work...