Loading...

A Lightweight Deep Learning Model for Online Network Traffic

Ameli, Ahmad Reza | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57165 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Jalili, Rasool
  7. Abstract:
  8. Traffic classification is crucial for the execution of many security and managerial tasks in a network, such as identifying malicious data, detecting disruptive users, and preventing the passage of traffic from certain applications. Nowadays, a significant portion of traffic in computer networks is encrypted. Therefore, the classification of encrypted traffic requires the use of solutions for analyzing encrypted traffic. In recent years, many solutions based on machine learning and deep learning have been proposed for analyzing encrypted traffic. While most of these solutions focus on improving the accuracy of network traffic classification, less attention has been paid to their computational speed and efficiency. However, computational speed and efficiency are of special importance for the practical and industrial use of these solutions. In this thesis, we aim to present a new method for compressing deep learning-based models for network traffic classification by examining existing deep learning-based models and the proposed solutions for model compression, and identifying their strengths and weaknesses, leading to a lightweight model with fast processing and high accuracy for network traffic classification. The proposed method in this thesis addresses model compression by leveraging the strengths of three existing solutions: iterative pruning, knowledge distillation, and weight quantization. The evaluation criteria for the proposed method include maintaining or increasing the model accuracy, reducing the model storage space, reducing the model inference time, and comparing the performance of the proposed method with the most prominent existing solutions in the literature, namely the lottery hypothesis and deep compression
  9. Keywords:
  10. Traffic Classification ; Machine Learning ; Deep Learning ; Pruning Method ; Compression ; Quantization

 Digital Object List

 Bookmark

No TOC