Machine Learning (ML) has become the shiny new object for security and is the foundational pillar of products such as Next-Generation Antivirus (NGAV) and User and Entity Behavior Analytics (UEBA). While most of these products have promised to be a “silver bullet” against malware, complete protection remains elusive. In fact, ML is more likely to detect and cure cancer than to stop all of today’s advanced threats for a number of reasons:
• The past doesn’t predict the future
• Nothing will keep the bad guys out
• The harder you try the more you fail
• You can’t always be connected
• It’s a black box
Shahid N. Shah, an internationally recognized cybersecurity and risk management expert, and Rene Kolga, Senior Director of Product Management at Nyotron, will explain these shortcomings and how to avoid them. Instead of chasing after an infinite number of malware variants and attack vectors, a different approach to malware detection is to focus on the finite intentions behind attacks, such as data exfiltration, corruption and deletion.