Detecting DoS Attacks through Synthetic User Behavior with Long Short-Term Memory Network.

Area of Science:

• Computer Science
• Cybersecurity
• Artificial Intelligence

Background:

• Modern Denial of Service (DoS) attacks are increasing in size and complexity.
• There is a growing need to understand and counter the role of Machine Learning (ML) in both executing and defending against these attacks.

Purpose of the Study:

• To investigate the potential of ML, specifically Long Short-Term Memory (LSTM) networks, in generating behavioral telemetry data.
• To explore the use of ML in creating spoofed network traffic that appears legitimate.

Main Methods:

• A custom testing environment was developed to capture and analyze user interactions, such as mouse and keyboard events.
• Long Short-Term Memory (LSTM) networks were employed to generate behavioral data.

Main Results:

• The study demonstrated the capability of ML models to generate synthetic behavioral telemetry data.
• The generated data can be used to create spoofed traffic, making attacks harder to detect.

Conclusions:

• While current economic factors limit the immediate threat of this ML-driven attack, future technological advancements may increase its feasibility.
• Continuous research and proactive development of countermeasures are essential to address evolving cyber threats and protect against sophisticated DoS attacks.