FedSSL-NTC: A Robust Federated Self-Supervised Learning Framework for Network Traffic Classification Under Privacy Constraints
Bibliographic record
Abstract
Network traffic classification (NTC) plays an essential role in managing, securing, and optimizing networks. Supervised learning methods face challenges such as label scarcity. Given that network traffic contains sensitive data and is distributed across multiple nodes, privacy-aware and scalable approaches are necessary for real-world deployment. In this paper, we introduce FedSSL-NTC, a privacy-enhancing federated framework that integrates self-supervised learning (SSL) and a traffic-adapted confident learning (CL) approach. In FedSSL-NTC, clients locally pretrain SSL models (Autoencoders or Tabular Contrastive Learning) and generate pseudo-labels. CL is then applied on the client side to reduce pseudo-label noise before federated classifier training. Robustness to non-independently and non-identically distributed (non-IID) data and class imbalance is achieved via FedProx, class-weighted loss, and a sample-size weighted FedAvg aggregation method. This framework uses Secure Aggregation to protect individual updates. On a self-generated + ISCX VPN-nonVPN dataset and the UCDavis–QUIC dataset, FedSSL-NTC achieves 95.88% and 98.24% accuracy (vs. centralized 96.29% and 98.76%), while reducing training time by approximately 4–5× through parallel client updates. The method outperforms recent federated/self-supervised baselines on the same evaluation protocol (e.g., 6% improvement compared to FS-GAN). Therefore, FedSSL-NTC offers a practical path to high-accuracy NTC under privacy constraints, non-IID distributions, and label scarcity.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.002 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.007 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".