Smart Email Security Assistant

artificial neural networks, email security, indicators of compromise, natural language processing, phishing

Authors

  • Cristian PASCARIU
    crpascariu@gmail.com (Primary Contact)
    Faculty of Electronics, Telecommunications and Information Technology, University POLITEHNICA of Bucharest, Romania
  • Ioan BACIVAROV EUROQUALROM, Faculty of Electronics, Telecommunications and Information Technology, University POLITEHNICA of Bucharest, Romania
2023-05-30

Downloads

With security incidents and breaches growing each year, email is still used as the major entry point to server malicious content that results in credential theft or malware infections enabling malicious threat actors to mount complex attacks. This paper is intended to document a new approach for detecting suspicious and malicious emails leveraging techniques such as security analytics, natural language processing to discover the intent of the email, as well as artificial neural networks to support more complex rules for classification. This solution can be used in a basic mode to flag which emails are safe and which are not, at the same time it can also be used by security analysts to gain a better understanding of the attack vectors and speed up the investigation process.