Multi-Objective Seagull Optimization Algorithm with Deep Learning-Enabled Vulnerability Detection for Secure Cloud...

Area of Science:

• Cloud computing security
• Cybersecurity
• Machine learning applications

Background:

• Cloud computing offers cost-effective services but faces significant security challenges.
• Intrusion detection systems (IDS) are crucial for identifying normal and anomalous network behaviors.
• Machine learning (ML) techniques enhance IDS by learning patterns and predicting outcomes.

Purpose of the Study:

• To design a novel multi-objective seagull optimization algorithm with deep learning-enabled vulnerability detection (MOSOA-DLVD) for cloud platform security.
• To enhance intrusion detection and classification accuracy in cloud environments.
• To improve the performance of deep belief network (DBN) algorithms through hyperparameter tuning.

Main Methods:

• Implemented a feature selection (FS) method using the multi-objective seagull optimization algorithm (MOSOA).
• Utilized a deep belief network (DBN) for intrusion detection and classification.
• Applied the sooty tern optimization algorithm (STOA) for hyperparameter tuning of the DBN to improve detection accuracy.

Main Results:

• The MOSOA-DLVD technique demonstrated high proficiency in identifying vulnerabilities and attacks within cloud infrastructure.
• Achieved a maximum intrusion detection accuracy of 99.34% on a benchmark IDS dataset.
• Outperformed recent methods in intrusion detection, establishing the model's effectiveness.

Conclusions:

• The proposed MOSOA-DLVD technique offers a robust solution for securing cloud platforms against cyber threats.
• The integration of MOSOA for feature selection and STOA for DBN hyperparameter tuning significantly boosts detection accuracy.
• The study validates the MOSOA-DLVD system's capability to enhance cloud security through advanced ML techniques.