BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20260901T031939EDT-436276evAS@132.216.98.100
DTSTAMP:20260901T071939Z
DESCRIPTION:\nColloque du DIRO\n\nProvably Secure Machine Learning\n\nThe w
 idespread use of machine learning systems creates a new class of computer 
 security vulnerabilities where\, rather than attacking the integrity of th
 e software itself\, malicious actors exploit the statistical nature of the
  learning algorithms. For instance\, attackers can add fake data (e.g. by 
 creating fake user accounts)\, or strategically manipulate inputs to the s
 ystem once it is deployed. So far\, attempts to defend against these attac
 ks have focused on empirical performance against known sets of attacks. I 
 will argue that this is a fundamentally inadequate paradigm for achieving 
 meaningful security guarantees. Instead\, we need algorithms that are prov
 ably secure by design\, in line with best practices for traditional comput
 er security. To achieve this goal\, we take inspiration from robust statis
 tics and robust optimization\, but with an eye towards the security requir
 ements of modern machine learning systems. Motivated by the trend towards 
 models with thousands or millions of features\, we investigate the robustn
 ess of learning algorithms in high dimensions. We show that most algorithm
 s are brittle to even small fractions of adversarial data\, and then devel
 op new algorithms that are provably robust. Additionally\, to accommodate 
 the increasing use of deep learning\, we develop an algorithm for certifia
 bly robust optimization of non-convex models such as neural networks.\n
DTSTART:20180216T153000Z
DTEND:20180216T163000Z
LOCATION:Room 3195\, CA\, Pav. André-Aisenstadt
SUMMARY:Jacob Steinhardt\, Stanford University
URL:https://www.mcgill.ca/mathstat/channels/event/jacob-steinhardt-stanford
 -university-285069
END:VEVENT
END:VCALENDAR