Background and Objective: The constant growth of invasions and information theft by using infected software has always been a problem. According to McAfee labs in 2020, on average, 480 new viruses are created each hour. The means of identifying such threats, categorizing and creating vaccines may not be that fast. Thanks to the increasing processing power and the popularity of artificial intelligence, it is now possible to integrate intelligence on an antivirus engine to enhance its protecting capabilities. And doing so with good algorithms and parameterization can be a key asset in securing one’s environment. In this work we analyze the overall performance of our authorial antivirus and compare it with other state-of-art antiviruses. The means of identifying such threats, categorizing and creating vaccines may not be that fast. Thanks to the increasing processing power of computers and the popularity of artificial intelligence, it is now possible to integrate intelligence on an antivirus engine to enhance its protecting capabilities. And doing so with good algorithms and parameterization can be a key asset in securing one’s environment. In this work we are analyzing the overall performance of our authorial antivirus and comparing it with other state-of-art antiviruses. Methods: In this work we are using mELM, a deep neural network which can perform quick training sections and have a satisfactory accuracy when classifying unknown files that may or may not be infected with Citadel. Our virus database is built with many examples of well-known infected files and our results are compared with 7 other intelligent antiviruses created by other authors. Results: Our antivirus achieves an overall performance of 91.33% when classifying benign and malware PE (portable executable) programs. Different initial conditions, learning functions and architectures of our antivirus are investigated in order to maximize their accuracy. Conclusions: In this work we have found that mELMs implementations are feasible and its performance can match state-of-art implementations. Its training and classification time is among the smallest found and the accuracy when detecting citadel-infected PEs is acceptable.

Background and Objective: The constant growth of invasions and information theft by using infected software has always beena problem. According to McAfee labs in 2020, on average, 480 new viruses are created each hour.The means of identifying such threats, categorizing and creating vaccines may not be that fast. Thanksto the increasing processing power and the popularity of artificial intelligence, it is now possible tointegrate intelligence on an antivirus engine to enhance its protecting capabilities. And doing so withgood algorithms and parameterization can be a key asset in securing one’s environment. In this workwe analyze the overall performance of our authorial antivirus and compare it with other state-of-artantiviruses. The means of identifying such threats, categorizing and creating vaccines may not be that fast.Thanks to the increasing processing power of computers and the popularity of artificial intelligence,it is now possible to integrate intelligence on an antivirus engine to enhance its protecting capabili-ties. And doing so with good algorithms and parameterization can be a key asset in securing one’senvironment. In this work we are analyzing the overall performance of our authorial antivirus andcomparing it with other state-of-art antiviruses. Methods: In this work we are using mELM, a deep neural network which can perform quicktraining sections and have a satisfactory accuracy when classifying unknown files that may or maynot be infected with Citadel. Our virus database is built with many examples of well-known infectedfiles and our results are compared with 7 other intelligent antiviruses created by other authors. Results: Our antivirus achieves an overall performance of 91.33% when classifying benign andmalware PE (portable executable) programs. Different initial conditions, learning functions and ar-chitectures of our antivirus are investigated in order to maximize their accuracy. Conclusions: In this work we have found that mELMs implementations are feasible and its per-formance can match state-of-art implementations. Its training and classification time is among thesmallest found and the accuracy when detecting citadel-infected PEs is acceptable.

No abstract