BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Know the Unknowns: Holistically Addressing Uncertainties and Distu
 rbances in Learning-Enabled Cyber-Physical Systems - Qi Zhu (Northeastern 
 University)
DTSTART:20220727T150000Z
DTEND:20220727T153000Z
UID:TALK177008@talks.cam.ac.uk
DESCRIPTION:Future cyber-physical systems will employ complex learning-bas
 ed sensing\, computation\, and communication components for their percepti
 on\, planning\, control\, and coordination\, and could operate in highly d
 ynamic and uncertain environment with safety and security assurance. To re
 alize this vision\, we have to better understand and address the challenge
 s from the &ldquo\;unknowns&rdquo\; &ndash\; the inherent uncertainties fr
 om system inputs and model inaccuracies\, the unexpected disturbances from
  component failures\, environment interferences\, and malicious attacks\, 
 and the lack of analyzability of neural network-based machine learning tec
 hniques. In this talk\, I will discuss these challenges and present some o
 f our recent work to address these challenges in a holistic manner\, inclu
 ding 1) end-to-end verification\, design and adaptation methods for ensuri
 ng robust and safe application of neural networks in perception and decisi
 on making\; and 2) cross-layer methods for mitigating execution disturbanc
 es (e.g.\, timing violations\, soft errors\, malicious attacks) using weak
 ly-hard paradigm.
LOCATION:Discussion Room\, Newton Institute
END:VEVENT
END:VCALENDAR