The Program
Day 1
22nd September
09:30 - 09:50
Welcome/ Introduction
09:50 - 10:35
Keynote 1 : Earth System Emulation: Calibration, exploration and understanding
Speaker : Duncan Watson-Parris
10:35 - 10:55
1 Contributed presentation
Towards Transformer-based Homogenization of Satellite Imagery for Landsat-8 and Sentinel-2
Speaker : Venkatesh Thirugnana Sambandham
Coffee break 10:55 - 11:25
11:25 - 12:10
Keynote 2 : Almost 5 years of deep learning in Earth system modelling -- where do we come from and where do we go?
Speaker : Peter DĂĽben
12:10 - 12:30
1 Contributed presentation
Transformers show high skill in forecasting atmospheric ozone pollution in Europe
Speaker : Sebastian Hickman
Lunch 12:30 - 13:30
13:30 - 14:10
2 Contributed presentations
Speaker : Matej Racic
Domain transfer experiments with attention based architectures for satellite image time series based crop type classification
Speaker : Samuel Barrett
14:10 - 14:55
Keynote 3 : Graph Neural Network & Transformers in Simulation
Speaker : Jonathan Godwin
14:55 - 15:40
3-minute poster introductions
Coffee break 15:40 - 16:00
16:00 - 17:30
Poster session
Workshop dinner 19:30
Day 2
23rd September
09:00 - 09:45
Keynote 4 : Learning and (re)using general visual representations
Speaker : Lucas Beyer
09:45 - 10:45
3 Contributed presentations
Speaker : Tony Boston
Towards Spatio-temporal Event Detection Using Transformer Models
Speaker : Diallo Diaoule
Atmorep: Representation Learning for Spatio-Temporal Data in the Earth Sciences
Speaker : Christian Lessig
Coffee break 10:55 - 11:25
11:15 - 11:25
Platforms, Datasets and Open Models
11:25 - 11:50
Plenary discussions on data and cloud platforms
11:50 - 12:30
2 Contributed presentations
Speaker : Tobias Finn
Multiscale Neural Operator: Learning Fast and Grid-independent PDE Solvers
Speaker : Bjoern Luetjens
Lunch 12:30 - 13:30
13:30 - 14:15
Speaker : Pedram Hassanzadeh
14:15 - 15:00
Plenary discussions on architectures and applications
15:00 - 15:15
Wrap-up and farewell
For an overview of all the posters, please refer to poster session section of home page.