Abstract— Recent years have seen a rise in cyber-attacks, which pose severe dangers to network security due to the expanding digital ecosystem. It is no longer possible to detect and prevent sophisticated cyber threats with only conventional security measures. Machine learning (ML) methods have shown promise as a means of bolstering network security in response to this problem. This paper provides an in-depth look at how different ML algorithms have been used to detect cyber assaults, and how that has the potential to improve network security. We investigate many forms of cyber attacks, list the drawbacks of traditional security methods, and examine how ML might be used to detect and counteract these dangers. The report includes a comprehensive analysis of the efficacy, advantages, and disadvantages of widely-used ML algorithms in network security, illuminating their applicability to various cyber attack scenarios. We also talk about the difficulties and potential of using ML to the field of network security. 

Keywords— Cyber Attacks, Network Security, Machine Learning, Intrusion Detection Systems (IDS), Intrusion Prevention Systems (IPS)