Unveiling the Source: Differentiating Human and Machine-Generated Texts in a Multilingual Setting
Bibliographic record
Abstract
Current advances in the text generative models and the increased use of these models in our daily lives have posed unique challenges to the text's authenticity. These machine-generated texts can be used in fake news, academic integrity violations, mimicking real personalities online, etc. Most existing works in this area have shown detection ability in monolingual environments that are not generalized whereas the latest generational models are proficient in multiple languages thus making it difficult for existing tools to perform detection effectively. We aim to bridge this gap in research by proposing two innovative approaches in multilingual settings: a Pretrained Language Model (PLM) based method and a stylometric feature-based technique. The former achieved a significant improvement over state-of-the-art techniques, demonstrating superior performance and generalizability. Concurrently, we also introduce a stylometric method that capitalizes on detailed textual features to outper-form traditional statistical and feature-based models, providing a valuable forensic analysis and authenticity verification tool. This method is efficient for low-resource languages and has low computing requirements. We evaluate these on a diverse multilingual dataset which demonstrates the high reliability and accuracy of our proposed models. This study not only advances the field of authorship attribution but also contributes to the development of more secure digital communication environments.
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 | 0.004 | 0.025 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.005 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.004 |
| Open science | 0.001 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.002 |
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".