BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Introduction to UQ - An Overview of the Challenges Posed by Comple
 x Spatio-Temporal Epidemic Models to inform policy (lessons learned so far
 ) - Daniel Williamson (University of Exeter)
DTSTART:20220208T133500Z
DTEND:20220208T140000Z
UID:TALK166801@talks.cam.ac.uk
DESCRIPTION:Uncertainty Quantification (UQ) is the subject devoted to quan
 tifying all of the uncertainties when you attempt to combine complex scien
 tific models and data to learn about the real world. We know that neglecti
 ng some of these uncertainties can bias inference\, throw off predictions 
 and is particularly problematic in decision making contexts. The UQ commun
 ity has been around since the early 1980's and a number of core methods an
 d tools exist for doing it. Fast forward to February and March 2020 and us
 ing models to inform decision makers entered the public conciousness like 
 never before as we used the best tools we had to understand and predict th
 e trajectory of COVID-19 in the UK. The UQ community waved the flag for th
 e importance of proper uncertainty quantification\, yet despite a number o
 f projects to synthesise UQ for COVID-19 modelling getting off the ground 
 through RAMP and later EPSRC\, UQ for COVID-19 never really broke through 
 into SPI-M/SAGE. In this talk I will overview some of the key ideas in UQ 
 and why they proved (and still prove) so difficult to implement for (our) 
 COVID-19 models. I'll view the problem through the lens of how we might be
  ready to include UQ right from the start of the next pandemic to help mod
 ellers better calibrate their models and report more accurate uncertaintie
 s to feed into policy support.&nbsp\;
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR