Classification of disinformation in hybrid warfare: an application of XLNet during the Russia’s war against Ukraine
Bibliographic record
Abstract
The spread of disinformation has become a critical component of hybrid warfare, particularly in Russia’s war against Ukraine, where social media serves as a battlefield for influence and propaganda. This study develops a comprehensive methodology for classifying disinformation in the context of hybrid warfare, focusing on Russia’s war against Ukraine. The objective of this study is to address the challenges of disinformation detection, particularly the increased spread of propaganda due to hybrid warfare. The study focuses on the use of transformer-based language models, specifically, XLNet, to classify multilingual, context-sensitive disinformation. The tasks of this study are to analyze current research and develop a methodology to effectively classify disinformation using the XLNet model. The proposed methodology includes several key components: data preprocessing to ensure quality, application of XLNet for training on diverse datasets, and hyperparameter optimization to handle the complexities of disinformation data. The study used datasets containing pro-Russian and neutral/pro-Ukrainian tweets, and the XLNet model demonstrated strong performance metrics, including high precision, recall, and F1-scores across different dataset sizes. Results showed that accuracy initially improved with increasing data volume but declined slightly with numerous datasets, suggesting the need for balancing data quality and quantity. The proposed methodology addresses the gaps in automated disinformation detection by integrating transformer-based models with advanced preprocessing and training techniques. This research improves the capacity for real-time detection and analysis of disinformation, thus contributing to public information governance and strategic communication efforts during wartime.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.003 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".