# Unknown-Traffic Detection, Calibration and Shortcut Reliance in Distilled Encrypted-Traffic Classifiers over One Year

arXiv: https://arxiv.org/abs/2609.31141 · HTML: https://mlsanity.com/paper/2609.31141/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Federated, robust and private distillation, one of 12 directions in Knowledge distillation.
- Builds on 43 of its 49 references that are in the map.
- First posted on arXiv 25 Sep 2026.

## Placed in

- Knowledge distillation / [Federated, robust and private distillation](https://mlsanity.com/distill/federated-robust-and-private-distillation/)

## Builds on (most cited first, 30 of 43)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Similarity-Preserving Knowledge Distillation](https://mlsanity.com/paper/1907.09682/) (2019)
- [Does Knowledge Distillation Really Work?](https://mlsanity.com/paper/2106.05945/) (2021)
- [Characterising Bias in Compressed Models](https://mlsanity.com/paper/2010.03058/) (2020)
- [Ensemble Distribution Distillation](https://mlsanity.com/paper/1905.00076/) (2019)
- [ET-BERT: A Contextualized Datagram Representation with Pre-training Transformers for Encrypted Traffic Classification](https://mlsanity.com/paper/2202.06335/) (2022)
- [What Knowledge Gets Distilled in Knowledge Distillation?](https://mlsanity.com/paper/2205.16004/) (2022)
- [Teacher's pet: understanding and mitigating biases in distillation](https://mlsanity.com/paper/2106.10494/) (2021)
- [Scaling Ensemble Distribution Distillation to Many Classes with Proxy Targets](https://mlsanity.com/paper/2105.06987/) (2021)
- [Fine-grained TLS services classification with reject option](https://mlsanity.com/paper/2202.11984/) (2022)
- [Robustness Challenges in Model Distillation and Pruning for Natural Language Understanding](https://mlsanity.com/paper/2110.08419/) (2021)
- [Self-Distribution Distillation: Efficient Uncertainty Estimation](https://mlsanity.com/paper/2203.08295/) (2022)
- [Towards Understanding Why Label Smoothing Degrades Selective Classification and How to Fix It](https://mlsanity.com/paper/2403.14715/) (2024)
- [What Do Compressed Multilingual Machine Translation Models Forget?](https://mlsanity.com/paper/2205.10828/) (2022)
- [Who Taught You That? Tracing Teachers in Model Distillation](https://mlsanity.com/paper/2502.06659/) (2025)
- [Credal Ensemble Distillation for Uncertainty Quantification](https://mlsanity.com/paper/2511.13766/) (2025)
- [Antidistillation Fingerprinting](https://mlsanity.com/paper/2602.03812/) (2026)
- [Soft Labeling Affects Out-of-Distribution Detection of Deep Neural Networks](https://mlsanity.com/paper/2007.03212/) (2020)
- [Logit-Based Ensemble Distribution Distillation for Robust Autoregressive Sequence Uncertainties](https://mlsanity.com/paper/2305.10384/) (2023)
- [When Teachers Mislead: Spurious-Signal-Aware On-Policy Distillation](https://mlsanity.com/paper/2608.03632/) (2026)
- [When Simple Model Just Works: Is Network Traffic Classification in Crisis?](https://mlsanity.com/paper/2506.08655/) (2025)
- [The Role of Teacher Calibration in Knowledge Distillation](https://mlsanity.com/paper/2508.20224/) (2025)
- [Knowledge Distillation Detection for Open-weights Models](https://mlsanity.com/paper/2510.02302/) (2025)
- [Preventing Shortcut Learning in Medical Image Analysis through Intermediate Layer Knowledge Distillation from Specialist Teachers](https://mlsanity.com/paper/2511.17421/) (2025)
- [Adaptive Label Smoothing for Out-of-Distribution Detection](https://mlsanity.com/paper/2410.06134/) (2024)
- [Trust the uncertain teacher: distilling dark knowledge via calibrated uncertainty](https://mlsanity.com/paper/2602.12687/) (2026)
- [ResAware: Cross-Environment Website Fingerprinting via Resource-Privileged Distillation](https://mlsanity.com/paper/2606.17462/) (2026)
- [MERLOT: A Distilled LLM-based Mixture-of-Experts Framework for Scalable Encrypted Traffic Classification](https://mlsanity.com/paper/2411.13004/) (2024)
- [Distillation-Enhanced Clustering Acceleration for Encrypted Traffic Classification](https://mlsanity.com/paper/2508.02282/) (2025)
- [A Functional Perspective on Knowledge Distillation in Neural Networks](https://mlsanity.com/paper/2510.12615/) (2025)

## Abstract (arXiv)

Knowledge distillation is the standard way to compress encrypted-traffic classifiers for the edge, and almost all such work judges students by accuracy alone. We ask what else a student inherits: unknown-traffic detection, calibration, shortcut reliance, and whether any survives a year of drift. Resemblance proves little on its own, since soft targets also regularise. We therefore distil one 101k-parameter student from two teachers of equal accuracy but different construction, a five-member ensemble and a single wider model, so that following one rather than the other is attributable to it. The design was pre-registered before any test result was seen. We tested ten hypotheses on CESNET-TLS-Year22, a year of real TLS traffic, across 18 test windows over 35 weeks. Two are supported: a student's per-flow unknown-scores shift toward its own teacher, but only at a conventional temperature, not the accuracy-optimal one; and a shortcut-reliant teacher passes its over-confidence to a student that never sees the feature. The drift prediction is reversed under both scores, the gap narrowing rather than widening and the student overtaking under the energy score in two of three replicates, as is the prediction that such a teacher harms its student's detection, which improves slightly. Shortcut reliance is set by model size, not distillation. Under the logit-based scores nothing else transfers: distillation beats neither a temperature-scaled direct student nor label smoothing. Exploratory analysis shows this turns on the scoring rule: with a feature-space detector the teacher detects unknown traffic 0.073 AUROC better than the direct student, where the energy score sees 0.000, and the conventional-temperature student inherits most of it. Label smoothing, with no teacher, recovers more. Distillation transfers the teacher's habits; what looks like an inherited ability is available without one.