BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Interpreting data-driven forecast systems - do they behave in phys
 ically sensible ways? - Rachel Furner | British Antarctic Survey/Universit
 y of Cambridge
DTSTART:20210223T110000Z
DTEND:20210223T123000Z
UID:TALK155392@talks.cam.ac.uk
CONTACT:Tudor Suciu
DESCRIPTION:There is growing interest and activity in producing data-drive
 n weather and climate forecasts - systems which have learnt on data from e
 ither observation of process based models\, and are then able to produce a
  forecast using machine learning and statistical methods. These methods sh
 ow promise\, but to date there is limited investigation into precisely wha
 t these models are learning\, and how they are making predictions. To impr
 ove confidence in these new methods\, we need to ‘open the black box’ 
 of data-driven methods and begin to understand which processes are being c
 aptured in these models\, and with this any limitations they have.\n\nIn t
 his work\, we focus on a very simple regression model for ocean temperatur
 e. We assess the sensitivity of the developed regressor to its inputs usin
 g two methods. Firstly\, we directly analyse the coefficients to give insi
 ght into how this particular model relies on the different input variables
 . Secondly\, we perform a series of withholding experiments - retraining a
  set of new regressors\, each with a single input variable withheld at the
  training stage. These experiments give us more general insight into what 
 is needed for a regression model to be able to make skilful predictions. W
 e then further analyse some of these experiments to see more precisely how
  model skill is impacted by certain input variables.\n\nOur results show t
 hat for this simple regression model\, the behaviour is much in line with 
 the physical understanding we have of the system\, indicating that the mod
 el is learning\, to some extent\, the physics of the system. This gives in
 creased confidence in our ability to one day use data-driven models alongs
 ide traditional process based models.
LOCATION:https://zoom.us/j/6708259482?pwd=Qk03U3hxZWNJZUZpT2pVZnFtU2RRUT09
END:VEVENT
END:VCALENDAR