Code for "CyberWallE at SemEval-2020 Task 11: An Analysis of Feature Engineering for Ensemble Models for Propaganda Detection" (V. Blaschke, M. Korniyenko & S. Tureski, 2020)

[EMNLP 2023] Official repository of paper titled "Detecting Propaganda Techniques in Code-Switched Social Media Text"

Fine-grained Propaganda Detection at NLP4IF 2019

This is the repo for the NLP4IF workshop shared task at EMNLP 2019

WANLP 2022: Multilingual Multi-Granularity Network for Propaganda Detection

DETECTION OF PROPAGANDA TECHNIQUES IN NEWS ARTICLES

An end-to-end pipeline for detecting propaganda in text. The model, built using GloVe word embeddings and a Multi-Layer Perceptron architecture, classifies sentences as "propaganda" or "non-propaganda". The project integrates training, model evaluation, serialization, and inference for streamlined deployment and usage.

Hierarchical, multilingual, multimodal detection of persuasion techniques in memes (SemEval‑2024 Task 4), with a custom hierarchical loss and strong H‑F1 across 4 languages.

Telegram posts from Russian news, misinformation, and propaganda channels made during the first weeks of the 2022 Russian invasion of Ukraine.

Единый Реестр Кремлеботов Ютуба

Computationally analyzing propaganda network created amid a global pandemic and attempting to influence perceptions at a global scale- Proof of Concept

LLM-based pipeline for detecting 23 propaganda and persuasion techniques in text. Built on DSPy + OpenAI, with consistency voting, an LLM judge, and span-level visualizations.

A machine learning project focused on detecting fake news by generating propaganda-style training data. It introduces the PropaNews dataset and uses techniques like natural language inference and self-critical training to improve detection of human-written disinformation.

Screen text scraping, AI text analysis and mouse automation.

AI Comment Moderator 2026 🛡️ | Auto-Detect & Remove Toxic Text 🤖

High-performance propaganda detection for SemEval-2020 Task 11. Includes SOTA NLP models, a real-time Chrome extension, and claim extraction tools.

Computational pipeline for detecting Russian foreign-broadcast propaganda and its amplification of social cleavages in Cyprus. Features multilingual NLP (XLM-RoBERTa, BERTopic, stanza), multi-source scraping (Telegram, Twitter/X, RT, Sputnik), and interrupted time series analysis.

NLP-based historical text analyzer with GUI and web interface. Supports similarity analysis, theme detection, and propaganda classification.