mlsanityData to 14 Oct 2026

Pruned Traffic Trees: Native Semantic Compression with a Protocol-Structured Model Family for Encrypted Traffic Classification

arXiv 2608.21874 ↗First posted 22 Aug 2026cs.NI cs.AIcited by 1

Not on the map. Why?

Abstract (arXiv)

Deep learning has achieved strong performance in encrypted traffic classification (ETC), yet its computational cost limits deployment on resource-constrained network devices such as routers and middleboxes. Existing compression methods mainly operate on weights, channels, hidden representations, or predictions, but do not explicitly determine which protocol fields and structural contexts should remain. We propose Pruned Traffic Trees (PTT), a three-level protocol-structured model family that treats native protocol structures as compression units. PTT-Full learns protocol-structured representations and field salience from complete Protocol Tree Graphs (PTGs), with flow-level self-supervised learning and protocol-presence-aware sparse execution. The learned salience and TopK+k closure construct Distilled PTGs (PTG-Ds) for PTT-Distilled, while PTT-Lite inherits this topology and reduces width through structure-aligned transfer and flow-level logits distillation. Under flow-disjoint and Strong Information Information (SII)-masked settings, PTT-Full achieves Macro-F1 scores of 0.9519 and 0.9416 on CSTNET-TLS1.3 and CipherSpectrum, while PTT-Lite retains 0.9325 and 0.9136 with 80.3% and 61.3% fewer parameters, 98.85% and 98.78% lower effective GFLOPs, and 8.75× and 8.46× CPU inference speedups. These results demonstrate that treating protocol structure itself as the compression object enables effective performance-efficiency trade-offs for lightweight ETC.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 19 of its 37 references that are in the map. First posted on arXiv 22 Aug 2026.

Builds on

19 of 37 references.

References · 19

  1. 2021Scaling Vision with Sparse Mixture of Expertscited by 747
  2. 2024A Survey on Mixture of Experts in Large Language Modelscited by 237
  3. 2017Deep Packet: A Novel Approach For Encrypted Traffic Classification Using Deep Learningcited by 70
  4. 2022ET-BERT: A Contextualized Datagram Representation with Pre-training Transformers for Encrypted Traffic Classificationcited by 66
  5. 2021Cold Brew: Distilling Graph Node Representations with Incomplete or Missing Neighborhoodscited by 58
  6. 2018Deep Fingerprinting: Undermining Website Fingerprinting Defenses with Deep Learningcited by 43
  7. 2023Knowledge Distillation on Graphs: A Surveycited by 38
  8. 2022T2-GNN: Graph Neural Networks for Graphs with Incomplete Features and Structure via Teacher-Student Distillationcited by 32
  9. 2022FreeKD: Free-direction Knowledge Distillation for Graph Neural Networkscited by 23
  10. 2017Automated Website Fingerprinting through Deep Learningcited by 21
  11. 2020SAIL: Self-Augmented Graph Contrastive Learningcited by 20
  12. 2022Compressing Deep Graph Neural Networks via Adversarial Knowledge Distillationcited by 18
7 more
  1. 2018Var-CNN: A Data-Efficient Website Fingerprinting Attack Based on Deep Learningcited by 15
  2. 2019Tik-Tok: The Utility of Packet Timing in Website Fingerprinting Attackscited by 15
  3. 2025SoK: Decoding the Enigma of Encrypted Network Traffic Classifierscited by 15
  4. 2022Boosting Graph Neural Networks via Adaptive Knowledge Distillationcited by 14
  5. 2022Alignahead: Online Cross-Layer Knowledge Extraction on Graph Neural Networkscited by 5
  6. 2026TrafficMoE: Heterogeneity-aware Mixture of Experts for Encrypted Traffic Classificationcited by 1
  7. 2026Treat Traffic Like Trees: A Semantic-Preserving Hierarchical Graph-Based Expert Framework for Encrypted Traffic Analysiscited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Unknown-Traffic Detection, Calibration and Shortcut Reliance in Distilled Encrypted-Traffic Classifiers over One Yearcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026TDDM-Melatt: A Decoupled Memory and Diffusion Framework for Generalizable Encrypted Traffic Classification
  2. 2024MERLOT: A Distilled LLM-based Mixture-of-Experts Framework for Scalable Encrypted Traffic Classification
  3. 2026Mean Masked Autoencoder with Flow-Mixing for Encrypted Traffic Classification
  4. 2026Multimodal Reasoning with LLM for Encrypted Traffic Interpretation: A Benchmark
  5. 2026NetMamba+: A Framework of Pre-trained Models for Efficient and Accurate Network Traffic Classification
  6. 2026MalMoE: Mixture-of-Experts Enhanced Encrypted Malicious Traffic Detection Under Graph Drift
  7. 2026Lightweight GenAI for Network Traffic Generation: Fidelity, Augmentation, and Classification
  8. 2026PCAP-LM: An LLM-Native Text Representation for TLS Bulk Traffic Analysis