Evaluation of a GAN-Based Method for Anomaly Detection in Open RAN Based on Experimental 5G Data
Bibliographic record
Abstract
Open RAN introduces substantial flexibility and openness to radio access networks, offering significant advantages while simultaneously increasing vulnerability to complex anomalies and security threats. Although a range of anomaly detection methods exists for other domains, few have been systematically assessed in the context of Open RAN.In this paper, we adapt, employ and analyze a generative adversary network (GAN) model originally designed for unsupervised anomaly detection in multivariate time series data, applying it to the Open RAN domain, using a publicly available experimental 5G data set collected from the OpenIreland 5G testbed. We demonstrate how the GAN model combines a generative approach, using Long Short-Term Memory (LSTM) networks for reconstruction, with a discriminative approach, also based on LSTM, for classification. This integration enables the model to effectively capture complex temporal correlations within the key performance indicators (KPIs) of Open RAN. This method yields an anomaly score, which merges reconstruction error with the discriminator’s output to identify anomalies in real-time.Our findings demonstrate that GAN achieves good accuracy in detecting malicious activities, showcasing good precision and recall performance. Furthermore, we discuss practical considerations for deploying GAN as a security xApp within the near-real-time RAN Intelligent Controller (RIC), addressing aspects such as latency, scalability, and integration with the O-RAN architecture.To the best of our knowledge, this is the first study to apply GAN-based anomaly detection on the publicly available dataset from the OpenIreland 5G testbed.
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 | 0.002 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| 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".