Authorship Forensics Portal
Bibliographic record
Abstract
This paper presents the research outcome, Authorship Forensics Portal, leveraging both Statistical Natural Language Processing (SNLP) and Convolutional Neural Networks (CNN) techniques to differentiate documents written by humans and ChatGPTs. The portal allows teachers to (1) upload labeled data that contains written text and its author; (2) configure parameters that are required for training models, e.g., 2-class (i.e., human and ChatGPT) or 3-class (i.e., human, ChatGPT 3.5, and ChatGPT as well as the train/test set split ratio, validation set ratio, and validation accuracy threshold for stopping the training process; (3) review the details of a trained model, e.g., the train/test set, the time spent, the prediction results like numbers, true positive, false positive, precision, recall, and f-value, etc.; (4) make their own trained models be private so only themselves can see and use or be public so other teachers can also see and use; and, (5) ask a chosen trained model for its opinion on whether a piece of text written by human or generative Al (e.g., ChatGPT for 2•class prediction and ChatGPT 3.5 or ChatGPT 4 for 3-class prediction). The results demonstrate a significant ability of the models to distinguish between human and Al-written text, with highest precision 0.9868 (Fo_5 score 0.9647) for the 2-class (human and ChatGPT) testing subset and highest precision 0.9875 (Fo.5 score 0.9753) for the 3-class (human, ChatGPT 3.5, and ChatGPT 4) testing subset.
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 | 0.007 | 0.037 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.006 | 0.004 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.005 | 0.010 |
| Open science | 0.002 | 0.007 |
| Research integrity | 0.003 | 0.002 |
| Insufficient payload (model declined to judge) | 0.069 | 0.071 |
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".