Lujo Bauer, CMU
From 12.00 until 13.30
At CNB/F/110 (Lunch) + CAB/F/100.9 (Seminar), ETH Zurich
Universitätstrasse 6, 8092 Zurich
Abstract:
The risks of cyber attacks are increasing: attackers are increasingly well organized and sophisticated; at the same time, more and more computer systems are accessible over the Internet to be attacked, including computer systems that operate critical infrastructure such as power generation and transportation. In this talk I'll describe several recent research results that show how we can use machine learning to lower the risk of cyber attacks, but also that our increasing reliance on machine learning brings about new risks. First, I will talk about using machine learning to help detect vulnerabilities in web applications vastly faster than using traditional methods. Then, I will describe how several practical uses of machine learning---face recognition and malware detection---are susceptible to new attacks.