Domain adaptive deep semi-supervised transfer learning for anomaly detection in OpenWiFi
Bibliographic record
Abstract
Evaluating service connectivity of OpenWiFi to endusers calls for meticulous analysis of traffic patterns to identify any unusual behavior. This work proposes an automated approach to annotate the partially labeled OpenWiFi data more effectively, enabling discrimination of anomalous behavior from normal traffic behavior. Our framework comprises feature selection, semi-supervised learning (SSL), deep semi-supervised transfer learning (DSSTL), and K-Means clustering analysis for forecasting pseudo labels on unlabeled data, utilizing limited labeled information to extract relevant Key Performance Indicators (KPIs). We utilized one-dimensional convolutional neural network (1D-CNN) and Long Short-Term Memory (LSTM), enhanced by semi-supervised learning (SSL) and unsupervised K-Means to generate pseudo labels as anomaly or normal for unlabeled data. DSSTL involves domain adaptation by combining SSL, transfer learning, and K-Means clustering algorithm, thereby generalizing the forecasting quality of pseudo labels over conventional SSL and clustering approach. Our findings showcase DSSTL-based 1D-CNN with Mean Square Error (MSE) loss outperforms the analysis for SSL and DSSTL-based LSTM in acquiring pseudo labels for all proportions of unlabeled data, as indicated by Calinski and Harabasz (CH) score. It is noticeable that DSSTL-based 1D-CNN with MSE achieves a CH score that increases by 94% corresponding to 25% over 30% unlabeled samples, regardless of the increasing trend in CH score with an increase in proportions of unlabeled samples.
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 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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".