Reliable News or Propagandist News? A Neurosymbolic Model Using Genre, Topic, and Persuasion Techniques to Improve Robustness in Classification

2026-04-02

AI Summary

This work addresses the limited generalizability of existing language model–based approaches to propaganda news detection, which are often hindered by biases in training data. To enhance robustness, the authors propose a neuro-symbolic hybrid model that, for the first time, integrates interpretable symbolic features—such as genre, topic, and persuasive techniques—with fastText text embeddings through a dedicated neuro-symbolic fusion architecture. Experimental results demonstrate that the proposed method outperforms purely text-based baselines in cross-source evaluation scenarios. Ablation studies and interpretability analyses further confirm the effectiveness and meaningful contribution of the introduced symbolic features to the model’s performance.

Abstract

Among news disorders, propagandist news are particularly insidious, because they tend to mix oriented messages with factual reports intended to look like reliable news. To detect propaganda, extant approaches based on Language Models such as BERT are promising but often overfit their training datasets, due to biases in data collection. To enhance classification robustness and improve generalization to new sources, we propose a neurosymbolic approach combining non-contextual text embeddings (fastText) with symbolic conceptual features such as genre, topic, and persuasion techniques. Results show improvements over equivalent text-only methods, and ablation studies as well as explainability analyses confirm the benefits of the added features. Keywords: Information disorder, Fake news, Propaganda, Classification, Topic modeling, Hybrid method, Neurosymbolic model, Ablation, Robustness

Problem

• Propaganda
• Fake news
• Classification
• Robustness
• Information disorder

Innovation

• Neurosymbolic model
• Propaganda detection
• Robustness
• Hybrid method
• Symbolic features