- 211 views
- 132 downloads
Analysis of Artificial Intelligence Techniques to Detect, Prevent, Analyze and Respond to Malware
-
- Author(s) / Creator(s)
-
Malware has been posing a significant problem to almost every organization. Every organization online and every user using the services of the internet is susceptible to attacks. The preparators of these attacks are humans; malware is the tool they use to exploit the systems. The “malicious” software is designed to carry out activities such as spying on user activity, encrypting critical information, making use of the host system’s applications without the user’s knowledge to access and steal critical data, creating entry and exit points for an attacker to enter and exit the victim’s environment as they please – creating backdoors. Malicious activity can be done using the applications on the system is not done without interacting with the operating system through the application programming interface (API) calls.
Many human resources are expended on monitoring systems to notice any anomalies caused by the running applications. It is a relatively slow process to detect such anomalies when the count of systems to be monitored on an organizational level increase from thousands to tens of thousands. Artificial intelligence (AI), a trendy industry buzzword, can expedite the detection of any malicious activity and assist IT professionals.
This report draws a basic understanding of malware and artificial intelligence concepts. It also presents how artificial intelligence can be used to detect malware. Anything detected can be prevented, analyzed, and responded to. In the lab implementation section, with the help of an AI model, it is demonstrated how artificial intelligence can detect malicious activity by malware through its invoked API calls. A sequential deep-neural network with a multi-class classification algorithm – softmax regression, is developed to classify malware based on its API calls.
Artificial intelligence serves as an excellent tool for detecting malware. AI's learning capabilities to learn from vast amounts of data make it a powerful tool. Especially in the cybersecurity industry, where the possibility of zero-day attacks is never nil, AI could help identify these
attacks by learning from all the previous data, recognizing patterns, and predicting the likelihood of one.
This report could serve as a reference to anyone seeking to develop an AI application to detect malware and educate one on malware and AI and AI detection capabilities. -
- Date created
- 2023-04-01
-
- Type of Item
- Report