BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Virtual Event: ECMWF-ESA Workshop on Machine Learning for Earth Sy
 stem Observation and Prediction
DTSTART:20201005T100000Z
DTEND:20201008T143000Z
DTSTAMP:20260415T003300Z
UID:indico-event-172@events.ecmwf.int
CONTACT:events@ecmwf.int
DESCRIPTION:\n\n#AIforEOWS\n\nWorkshop motivation and description\n\nMachi
 ne Learning/Deep Learning (ML/DL) techniques have made remarkable advances
  in recent years in a large and ever-growing number of disparate applicati
 on areas\, e.g. natural language processing\, computer vision\, autonomous
  vehicles\, healthcare\, finance and many others. These advances have been
  driven by the huge increase in available data\, the increase in computing
  power and the emergence of more effective and efficient algorithms.\n\nEa
 rth System Observation and Prediction (ESOP) have arguably been latecomers
  to the ML/DL party\, but interest is rapidly growing\, and innovative app
 lications of ML/DL tools are also becoming increasingly common in ESOP.\n\
 nThe interest of ESOP scientists in ML/DL techniques stems from different 
 perspectives. From the observation side\, the current and future availabil
 ity of satellite-based Earth System measurements at high temporal and spat
 ial resolutions and the emergence of entirely new observing systems made p
 ossible by ubiquitous internet connectivity (so called “Internet Of Thin
 gs”) pose new challenges to established processing techniques and ultima
 tely to our ability to make effective use of these new sources of informat
 ion. ML/DL tools can potentially be useful to overcome some of these probl
 ems\, for example in the areas of observation quality control\, observatio
 n bias correction and the development of efficient observation operators a
 nd observation-based retrievals.\n\nFrom a data assimilation perspective\,
  ML/DL approaches are interesting because they can be typically framed as 
 Bayesian inference problems using a similar methodological toolbox as the 
 one used e.g. in variational data assimilation. It can be argued that some
  of the techniques already common in the data assimilation community (e.g.
  model error estimation\, model parameter estimation) are effectively a ty
 pe of ML/DL. The question is then\, what lessons can the ESOP community le
 arn from the methodologies and practices of the ML/DL community? Can we se
 amlessly integrate these new ideas into current data assimilation practice
 s?\n\nML/DL solutions are also being explored for model identification\, e
 ither in terms of the full forecast model or for specific model parametriz
 ations which are computationally expensive and/or physically uncertain. Ho
 w to best combine physical knowledge with the statistical knowledge provid
 ed by ML/DL approaches is an important and open question. Various types of
  machine learning technologies have also a rather long history of applicat
 ion in model interpretation and post-processing. The question of how ML/DL
  can help us extract more value from environmental forecasts is thus a rel
 evant and current one to pose.\n\nAn important issue are the uncertainty c
 haracteristics of the ML results\, and to understand better what physical 
 relations they have been trained on. Many methodologies for both uncertain
 ty quantification and for back-tracing ML output to input features have be
 en proposed\, but there is not yet a consensus view. Progress here is need
 ed to improve and better understand reliability of ML results\, which is c
 rucial in an operational context.             \n\nWorkshop aim
 s\n\nIn the application of ML/DL techniques to ESOP there are still many u
 nanswered questions. The aim of the workshop was to appraise the state of 
 the art of the application of ML/DL techniques to ESOP\, to identify the m
 ain issues that need to be solved for further progress\, and to make a sta
 rt on charting ways forward. Presenters of the longer talks covered not ju
 st their own work but also to gave a general overview of the subject. Disc
 ussions were facilitated by parallel working groups where the main issues 
 were discussed in more detail. The output of the workshop is in the form o
 f working group reports\, to be summarised in a technical memorandum or pa
 per.\n\n\n\n\n \n\nhttps://events.ecmwf.int/event/172/
IMAGE;VALUE=URI:https://events.ecmwf.int/event/172/logo-3604055506.png
LOCATION:ECMWF
URL:https://events.ecmwf.int/event/172/
END:VEVENT
END:VCALENDAR