BEGIN:VCALENDAR
PRODID:-//Columba Systems Ltd//NONSGML CPNG/SpringViewer/ICal Output/3.3-
 M3//EN
VERSION:2.0
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTAMP:20241209T081334Z
DTSTART:20250212T140000Z
DTEND:20250212T150000Z
SUMMARY:AI Security: Language Models\, Data Encryption\, Software Verific
 ation
UID:{http://www.columbasystems.com/customers/uom/gpp/eventid/}e19f-m4gra0
 1s-er4cd1
DESCRIPTION:Neural networks are slowly getting integrated into safety-cri
 tical systems. Unfortunately\, we still lack a full suite of algorithms 
 and tools to guarantee their safety. In this talk\, I will present a few
  open challenges in AI safety and security: consistent behaviour in lang
 uage models\, machine learning over encrypted data\, model compression w
 ith error guarantees\, bug-free floating-point software. Here\, I will c
 laim that formal methods are the key to address these challenges\, as lo
 ng as we can settle on an unambiguous specification.\n\n
STATUS:TENTATIVE
TRANSP:TRANSPARENT
CLASS:PUBLIC
LOCATION:Kilburn Building\, Manchester
END:VEVENT
END:VCALENDAR