TAVo: Tor Application Detection with Voting Critic
Bibliographic record
Abstract
The Onion Router (TOR) Network, experiences an increasing number of users with an even more rapid increase in the volume of data usage. This leads to network congestion as well as a general decrease in end-user performance. TOR network is one of the solutions adopted by attackers that allows them to hide their identity through encryption. In this scenario, monitoring the applications and their traffic is a challenging task that can however bring relevant insights to network administrators and security teams. In this paper, we present a specially designed system to analyze the packet capture traffic of a network and classify each stream based on whether it is using the TOR network or not, as well as classifying the type of application using the TOR network. TOR Application detection with a Voting-classifier Critic (TAVo) uses a two-layer classifier that is combined with a specialized Critic model. The first layer aims at separating the Non-TOR from the TOR traffic, while the second layer determines the application that caused the TOR traffic. Whenever the second layer outputs a prediction with lower confidence the voting-classifier Critic is called to confirm or correct the predictions of these difficult cases. Through a set of experiments on a recent dataset, we show that TAVo has important advantages in terms of performance, achieving an average F1 score of 84% without the Critic model, and 91% with the Critic. Moreover, because the Critic model is only used for the cases where the base models face more difficulties, the overall system is efficient.
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 | 0.003 | 0.008 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.002 |
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".