Mihail Plesa works as a Security Software Engineer at Special Telecommunications Service. Mihail holds an MSc in Information Security from Military Technical Academy and he is currently pursuing a PhD in the field of applied cryptography at the Faculty of Mathematics and Informatics at the University of Bucharest. He is passionate about applying cryptography to solve problems ranging from basic confidentiality to privacy-preserving computation. Mihail has authored and co-authored research papers on topics like privacy-preserving machine learning or quantum cryptography.

Augment cybersecurity through A.I.

We are all aware of the recent advances in Artificial Intelligence. A.I. can generate images from text, discover more efficient algorithms, and even solve problems from the International Math Olympiad faster than any participant.

What is probably not widely known are all the cool things AI can do in cybersecurity.

In this talk, we will look at how A.I. can solve different problems in cybersecurity. We will see how to use machine learning to perform malware analysis tasks, social engineering, or penetration testing.

Some specific topics we will discuss include:

• Detect obfuscated JavaScript files
• Perform malware analysis using Convolutional Neural Networks (CNNs)
• Generate new malware using Generative Adversarial Neural Networks (GANs)
• Create a Twitter phishing boot
• Perform voice impersonation
• Using Metasploit with deep neural networks to accelerate the process of penetration testing