Incremental Learning for Mobile Encrypted Traffic Classification
Инкрементное обучение для классификации зашифрованного мобильного трафика
2021-06-01
SCID: 54.1/23wd7vjb
Discuss with AI
catastrophic forgettingincremental learningmobile encrypted traffic classificationone-vs-rest strategysample selection
Figures from the paper
Abstract (AI)
With the rising popularity of mobile networks and applications, network traffic classification has gradually become essential to mobile network management and cyberspace security. Existing state-of-the-art methods have achieved high accuracy in the closed-world mobile encrypted traffic classification, where the classifier only needs to process the classes seen in the training. When we update the dataset with new mobile applications, these methods must retrain a new classifier from scratch to learn the knowledge of all applications because directly fine-tuning the existing classifier would lead to the catastrophic forgetting problem. Thus, it is challenging to incrementally add new applications to the classification system while preserving the learned knowledge of the existing classifier. To tackle this issue, we propose an incremental learning framework based on the one vs rest (OvR) strategy and neural network classifiers. Moreover, we adopt a sample selection algorithm to balance the conflict between the growing training effort caused by new applications and the high classification accuracy. The experimental results demonstrate that our proposed framework achieves incremental learning with high classification accuracy like the closed-world method, and the selection algorithm significantly reduces training efforts to meet the dataset scale control and classification accuracy requirement in the lifetime incremental learning.
Key Findings
1
A sample-selection algorithm balances increasing training costs from new applications against classification accuracy and dataset-size constraints.
2
Experiments show the framework achieves high incremental classification accuracy comparable to closed-world methods.
3
It proposes an incremental learning framework combining one-versus-rest classification with neural-network classifiers to mitigate catastrophic forgetting.
4
The paper addresses incremental mobile encrypted traffic classification while preserving knowledge of previously learned applications when new classes arrive.
5
The selection algorithm significantly reduces training effort while satisfying dataset-scale and accuracy requirements during lifetime incremental learning.
Research Object
mobile encrypted network traffic from mobile applications
Research Subject
incremental classification performance, knowledge retention, and training-effort control when adding new application classes
Publication Details
Publication Date
2021-06-01
Journal
Publisher
ISSN
Cited by
32
Access Type
Author Information
Download PDF
Subscribe to digest