Artificial Reality | The PatternEx Blog

AI For Enterprise Security: The Challenges from a Data Scientist's Perspective

Looking for a overview of cybersecurity and artificial intelligence? Look no further. In this 45 minute session delivered at the Center for Long Term Cybersecurity (UC-Berkeley), Dr. Ignacio Arnaldo shares his view at the intersection of Big Data, InfoSec, and Artificial Intelligence. 


Using Computer Vision as an analogy, Dr. Arnaldo will teach you how AI works, the challenges it faces in cybersecurity, and the transformative potential it has. 

This presentation was given May 2, 2017 at the Center for Long Term Cybersecurity at UC-Berkeley.

Content and time:

  • 2:30 - Intro of AI and InfoSec
  • 8:20 - Primer on AI
    • Feature extraction, Labelling, Training new Machine Learning Models
    • Computer Vision example and contrast with InfoSec
    • Key challenges for Data Science in InfoSec
  • 13:54 - Advanced Persistent Threats / Advanced Targeted Attacks
  • 19:05 - Summary of Challenges and Introduction to Solutions
    • Big Data Pipelines
    • Active Learning
  • 25:08 - Research / IEEE Paper Results
  • 26:00 - How to effectively engage humans to train AI
  • 30:00 - Sharing attack patterns and sharing across multiple organizations
    • Transfer Learning
  • 33:05 - Addressing the variety of threats with AI
  • 40:00 - PatternEx background

From the Center for Long Term Security's site:

In his presentation, Dr. Arnaldo provided an overview of how machine learning can support cybersecurity. He drew parallels with computer vision to explain how machines learn (and why cybersecurity is a unique machine-learning challenge), and he explained that the evolution of artificial intelligence in the field cybersecurity parallels that of a cybersecurity analyst, who learns through experience to master different phases, including building knowledge, detection, investigation, and response.

The link to the presentation is here:


Topics: Ignacio Arnaldo CLTC