Predicting Malware Attacks with The Help of Machine Learning
Disciplines
Artificial Intelligence and Robotics | Other Computer Sciences
Abstract (300 words maximum)
Malware has become a more widespread problem alongside the rapid growth of technology. Malware is any unwanted software placed in the system without knowledge or consent for performing malicious activities. Despite the implementation and adoption of novel preventative procedures, the number of malware attacks steadily rises every year. These attacks can spread through networks and disrupt any operation, leading to many cyber thefts, manipulated data, and most often than not, ransom. Unfortunately, many of the anti-malware systems that are available today are only able to defend the system from malware that is already identified. To address this problem for this research, we constructed deep learning models to estimate the likelihood of a computer will become infected with malicious software. Deep neural networks can process data in a way that works similar to the human brain with very little human assistance to learn relationships between complex data. We envision that with the assistance of our deep learning model, we will be able to reduce the number of compromised systems significantly by predicting malicious activity before it happens.
Academic department under which the project should be listed
CCSE - Computer Science
Primary Investigator (PI) Name
Md Abdullah Al Hafiz Khan
Predicting Malware Attacks with The Help of Machine Learning
Malware has become a more widespread problem alongside the rapid growth of technology. Malware is any unwanted software placed in the system without knowledge or consent for performing malicious activities. Despite the implementation and adoption of novel preventative procedures, the number of malware attacks steadily rises every year. These attacks can spread through networks and disrupt any operation, leading to many cyber thefts, manipulated data, and most often than not, ransom. Unfortunately, many of the anti-malware systems that are available today are only able to defend the system from malware that is already identified. To address this problem for this research, we constructed deep learning models to estimate the likelihood of a computer will become infected with malicious software. Deep neural networks can process data in a way that works similar to the human brain with very little human assistance to learn relationships between complex data. We envision that with the assistance of our deep learning model, we will be able to reduce the number of compromised systems significantly by predicting malicious activity before it happens.