Back To Schedule
Saturday, October 20 • 10:00am - 10:30am
Adversarial machine learning to detect phishing URLs

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Feedback form is now closed.
Phishing attacks cost businesses millions of dollars and are a principal vector for data breaches. Attackers and defenders are using artificial intelligence in an escalating arms race that threatens to consume ever greater resources. It does not have to be this way. This presentation describes the use of practical, cost-effective machine learning tools to generate and to detect target phishing URLs.
Three developments help bring phishing prevention under control. First, many machine learning platforms give essentially the same answer to the problems that they are given, so neither the offense nor the defense needs to know precisely what the other side is doing to be effective. Second, the use of adversarial training methods means that defenders can anticipate the future challenges of phishing attempts. Third, relatively simple, cost effective methods can be used generate novel phishing attempts and to block them, reducing the cost to defend against future phishing attacks while improving the defense effectiveness.

avatar for Herbert Roitblat

Herbert Roitblat

Principal Data Scientist, Mimecast
Dr. Roitblat, is currently Principal Data Scientist, for Mimecast where he works on data analysis and visualization, information retrieval, machine learning, and computer security. Before joining Mimecast, Dr. Roitblat was the Principal Data Scientist for Archiving and eDiscovery... Read More →

Saturday October 20, 2018 10:00am - 10:30am
Ballroom # 406