Loading...

Anomaly Based Intrusion Detection in Computer Networks Using Generative Adversarial Networks

Heidary, Milad | 2020

694 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53615 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Hemmatyar, Ali Mohammad Afshin
  7. Abstract:
  8. Due to the rapid development of computer networks, security is a major concern. Methods of intruding computer networks are also rapidly developing, and there is a new method every day. These facts corroborate the need for new and more intelligent mechanisms for detecting intrusion. To detect intrusion, one must analyze the network traffic. The most used traditional methods of traffic separation are port-based and payload based detection. The former is not so efficient, and the latter is not only inefficient but also violates the privacy of users. Unsatisfied by such methods, researchers adopted machine learning techniques and tried to develop new solutions for detecting intrusion. Methods developed so far perform better than traditional ones but still are not good enough for modern networks. Considering the vast advances in machine learning techniques, including deep learning, leading to even better methods. The problem is that these new methods have high false-positive rates. In this research, we will propose and evaluate a solution based on a new technique in machine learning called generative adversarial networks. This method is accurate enough and also has a low false-positive rate
  9. Keywords:
  10. Computer Networks ; Generative Adversarial Networks ; Anomaly Detection ; Intrusion Detection System ; Network Security

 Digital Object List

 Bookmark

...see more