BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Aardvark weather: end-to-end data-driven weather forecasting - Ric
 hard Turner\; University of Cambridge
DTSTART:20241119T130000Z
DTEND:20241119T140000Z
UID:TALK224281@talks.cam.ac.uk
CONTACT:Dr H Ge
DESCRIPTION:Weather forecasting is critical for a range of human activitie
 s including transportation\, agriculture\, industry\, as well as the safet
 y of the general public. Over the last two years\, machine learning models
  have shown that they have the potential to transform the complex weather 
 prediction pipeline\, but current approaches still rely on numerical weath
 er prediction (NWP) systems\, limiting forecast speed and accuracy. In thi
 s talk\, I will give some of the background on these developments. I will 
 then introduce a machine learning model which can replace the entire opera
 tional NWP pipeline. Aardvark Weather\, an end-to-end data-driven weather 
 prediction system\, ingests raw observations and outputs global gridded fo
 recasts and local station forecasts. Further\, it can be optimised end-to-
 end to maximise performance over quantities of interest. I will show that 
 the system outperforms an operational NWP baseline for multiple variables 
 and lead times for gridded and station forecasts. These forecasts are prod
 uced with a remarkably simple neural process model using just 8% of the in
 put data and three orders of magnitude less compute than existing NWP and 
 hybrid AI-NWP methods. We anticipate that Aardvark Weather will be the sta
 rting point for a new generation of end-to-end machine learning models for
  medium-range forecasting.
LOCATION:CBL Seminar room BE4-38
END:VEVENT
END:VCALENDAR