Virtual Event: ECMWF-ESA Workshop on Machine Learning for Earth System Observation and Prediction
Monday, 5 October | |||
→ |
Welcome and introduction - ECMWF
Speaker:
Andy Brown
(ECMWF)
|
ECMWF
Virtual
|
15m |
→ |
Welcome and introduction - ESA
Speaker:
Pierre-Philippe Mathieu
(ESA)
|
ECMWF
Virtual
|
15m |
Session 1: ML for the Earth System - Setting the scene
Chair:
Massimo Bonavita
(ECMWF)
|
|||
→ |
Accelerating and Explaining Earth System Process Models with Machine Learning
Speaker:
David Gagne
(National Center for Atmospheric Research)
|
ECMWF
Virtual
|
45m |
→ |
Learning from earth system observations: machine learning or data assimilation?
Speaker:
Alan Geer
(ECMWF)
|
ECMWF
Virtual
|
45m |
→ |
Lunch break
|
ECMWF
Virtual
|
1h |
Session 2: ML for Earth System Observations
Chair:
Bertrand Le Saux
(ESA)
|
|||
→ |
The Rise of AI for EO
Speaker:
Pierre-Philippe Mathieu
(ESA)
|
ECMWF
Virtual
|
30m |
→ |
Leveraging Modern AI/ML Techniques in NWP Including Data Fusion/Assimilation
Speaker:
Sid Boukabara
(NOAA)
|
ECMWF
Virtual
|
30m |
→ |
Exploring the Frontiers of Deep Learning for Earth System Observation and Prediction
Speaker:
David Hall
(NVIDIA)
|
ECMWF
Virtual
|
30m |
→ |
Coffee break
|
ECMWF
Virtual
|
30m |
Session 2 (cont.): ML for Earth System Observations
Chair:
Peter Dueben
(ECMWF)
|
|||
→ |
On the Interpretation of Neural Networks Trained for Meteorological Applications
Speaker:
Imme Ebert-Uphoff
(Colorado State University)
|
ECMWF
Virtual
|
30m |
→ |
Deep Hashing for Scalable Remote Sensing Image Retrieval in Large Archives
Speaker:
Begüm Demir
(TU Berlin)
|
ECMWF
Virtual
|
30m |
→ |
Poster session
|
1h 30m | |
Tuesday, 6 October | |||
Session 2 (cont.): ML for Earth System Observations
Chair:
Susanne Mecklenburg
(ESA)
|
|||
→ |
Using Machine Learning to advance hour-scale heavy rain forecast with high resolution ECMWF Global Model and Local Meso-scale Model Forecasts
Speaker:
Qi Zhong
(China Meteorological Administration Training Centre)
|
ECMWF
Virtual
|
30m |
→ |
Artificial Intelligence in ESA Contribution to the United Nations Framework Convention on Climate Change (UNFCCC) : An Overview
Speakers:
Eduardo Pechorro
(ESA Climate Office),
Amy Campbell
(National Oceanography Centre Graduate School)
|
ECMWF
Virtual
|
30m |
→ |
Neural network products in land surface data assimilation
Speaker:
Peter Weston
(ECMWF)
|
ECMWF
Virtual
|
30m |
→ |
Coffee break
|
ECMWF
Virtual
|
30m |
Session 3: ML for Data Assimilation
Chair:
Alan Geer
(ECMWF)
|
|||
→ |
Big Data, Big Computation, and Machine Learning in Numerical Weather Prediction
Speaker:
Takemasa Miyoshi
(RIKEN)
|
ECMWF
Virtual
|
30m |
→ |
Artificial Neural Network at the service of Data Assimilation (and vice versa)
Speaker:
Rossella Arcucci
(Imperial College London)
|
ECMWF
Virtual
|
30m |
→ |
Using machine learning and data assimilation to learn both dynamics and state
Speaker:
Marc Bocquet
(Ecole des Ponts ParisTech)
|
ECMWF
Virtual
|
30m |
→ |
Combining data assimilation and machine learning to emulate hidden dynamics and to infer unresolved scale pametrisation.
Speaker:
Alberto Carrassi
(University of Reading and NCEO (UK); University of Utrecht (NL))
|
ECMWF
Virtual
|
30m |
→ |
Lunch
|
ECMWF
Virtual
|
1h |
Session 3 (cont.): ML for Data Assimilation
Chair:
Peter Lean
(ECMWF)
|
|||
→ |
Data Assimilation and Machine Learning Science at ECMWF
Speaker:
Massimo Bonavita
(ECMWF)
|
ECMWF
Virtual
|
30m |
→ |
Joint learning of variational data assimilation models and solvers
Speaker:
Ronan Fablet
(IMT Atlantique)
|
ECMWF
Virtual
|
30m |
→ |
Hyperparameter learning in data assimilation systems
Speaker:
Manuel Pulido
(Universidad Nacional del Nordeste)
|
ECMWF
Virtual
|
30m |
→ |
Coffee break
|
ECMWF
Virtual
|
30m |
Session 3 (cont.) and Session 4: ML for Data Assimilation and ML for Product Development
Chair:
Marcin Chrust
(ECMWF)
|
|||
→ |
Bayesian Deep Learning for Data Assimilation
Speaker:
Peter Jan van Leeuwen
(Colorado State University)
|
ECMWF
Virtual
|
30m |
→ |
Machine Learning for Applied Weather Prediction
Speaker:
Sue Ellen Haupt
(National Center for Atmospheric Research)
|
ECMWF
Virtual
|
30m |
→ |
Deep Unsupervised Learning for Climate Informatics
Speaker:
Claire Monteleoni
(University of Colorado)
|
ECMWF
Virtual
|
30m |
→ |
Deep Learning for Post-Processing Ensemble Weather Forecasts
Speaker:
Nikoli Dryden
(ETH Zurich)
|
ECMWF
Virtual
|
30m |
Wednesday, 7 October | |||
Session 4 (cont.): ML for Product Development
Chair:
Bertrand Le Saux
(ESA/ESRIN)
|
|||
→ |
Automatic detection of weather events in high-resolution NWP outputs
Speaker:
Laure Raynaud
(Météo-France)
|
ECMWF
Virtual
|
30m |
→ |
Machine learning at ECMWF
Speaker:
Peter Dueben
(ECMWF)
|
ECMWF
Virtual
|
30m |
→ |
Probabilistic Deep Learning for Postprocessing Wind Forecasts in Complex Terrain
Speaker:
Daniele Nerini
(MeteoSwiss)
|
ECMWF
Virtual
|
30m |
→ |
Coffee break
|
ECMWF
Virtual
|
30m |
Session 4 (cont.) and Session 5: ML for Product Development and ML for Model Identification and development
Chair:
Sveinung Loekken
(ESA)
|
|||
→ |
Significance-tested and physically constrained interpretation of a deep-learning model for tornadoes
Speaker:
Ryan Lagerquist
(Cooperative Institute for Research in the Atmosphere (CIRA))
|
ECMWF
Virtual
|
30m |
→ |
Neural Networks for Postprocessing Ensemble Weather Forecasts
Speaker:
Sebastian Lerch
(Karlsruher Institut für Technologie (KIT))
|
ECMWF
Virtual
|
30m |
→ |
Towards an end-to-end data-driven weather model
Speaker:
Duncan Watson-Parris
(University of Oxford)
|
ECMWF
Virtual
|
30m |
→ |
Emulation of gravity wave parameterisation in weather forecasting
Speaker:
Matthew Chantry
(University of Oxford)
|
ECMWF
Virtual
|
30m |
→ |
Lunch break
|
ECMWF
Virtual
|
1h |
Session 5 (cont.): ML for Model Identification and development
Chair:
Massimo Bonavita
(ECMWF)
|
|||
→ |
Physics Guided Machine Learning: A New Framework for Accelerating Scientific Discovery
Speaker:
Vipin Kumar
(University of Minnesota)
|
ECMWF
Virtual
|
30m |
→ |
Assessing Machine Learning Approaches for Physical Parameterizations in Atmospheric General Circulation Models
Speaker:
Christiane Jablonowski
(University of Michigan)
|
ECMWF
Virtual
|
30m |
→ |
Model optimization with a genetic algorithm
Speaker:
Pieter Houtekamer
(Environment and Climate Change Canada)
|
ECMWF
Virtual
|
30m |
→ |
Coffee break
|
ECMWF
Virtual
|
30m |
Session 5 (cont.): ML for Model Identification and development
Chair:
Peter Jan van Leeuwen
(Colorado State University)
|
|||
→ |
Analog Forecasting of Extreme‐Causing Weather Patterns Using Deep Learning
Speaker:
Ashesh Chattopadhyay
(Rice University)
|
ECMWF
Virtual
|
30m |
→ |
What If The Easiest Part of the Global Atmospheric System For Machines To Learn Is The Dynamics?
Speaker:
Dale Durran
(University of Washington)
|
ECMWF
Virtual
|
30m |
→ |
S2S forecasting using large ensembles of data-driven global weather prediction models
Speaker:
Jonathan Weyn
(University of Washington)
|
ECMWF
Virtual
|
30m |
→ |
Description of Working Groups
Speaker:
Massimo Bonavita
(ECMWF)
|
ECMWF
Virtual
|
30m |
Thursday, 8 October | |||
Working groups
|
|||
→ |
WG1 (Observations) chaired by Alan Geer (ECMWF) and Bertrand Le Saux (ESA/ESRIN)
|
ECMWF
Virtual
|
1h 30m |
→ |
WG2 (Data Assimilation) chaired by Alberto Carrassi (Univ. of Reading) and Rosella Arcucci (Imperial College)
|
ECMWF
Virtual
|
1h 30m |
→ |
WG3 (Models) chaired by Massimo Bonavita (ECMWF) and Peter Dueben (ECMWF)
|
ECMWF
Virtual
|
1h 30m |
→ |
WG4 (Ensembles, Product Generation) chaired by Laure Raynaud (Météo-France) and Nicolas Longepe (ESA)
|
ECMWF
Virtual
|
1h 30m |
→ |
Coffee break
|
ECMWF
Virtual
|
30m |
Working groups
|
|||
→ |
WG1 (Observations) chaired by Alan Geer (ECMWF) and Bertrand Le Saux (ESA/ESRIN)
|
ECMWF
Virtual
|
1h 30m |
→ |
WG2 (Data Assimilation) chaired by Alberto Carrassi (Univ. of Reading) and Rosella Arcucci (Imperial College)
|
ECMWF
Virtual
|
1h 30m |
→ |
WG3 (Models) chaired by Massimo Bonavita (ECMWF) and Peter Dueben (ECMWF)
|
ECMWF
Virtual
|
1h 30m |
→ |
WG4 (Ensembles, Product Generation) chaired by Laure Raynaud (Météo-France) and Nicolas Longepe (ESA)
|
ECMWF
Virtual
|
1h 30m |
→ |
WG chairs finalise reports
|
30m | |
→ |
Lunch break
|
ECMWF
Virtual
|
1h |
→ |
Working Groups plenary discussion and close
Chair:
Massimo Bonavita
(ECMWF)
|
1h 30m |