资源环境科技发展态势分析平台

The third edition of the ECMWF–ESA Workshop on Machine Learning for Earth Observation and Prediction will take place from 14 to 17 November 2022. Registration and abstract submission are now open.

The event aims to provide an up-to-date snapshot of the state of the art in this rapidly evolving field and to facilitate discussion among scientists and practitioners.

“The workshop will bring together Earth science and artificial intelligence/machine learning (AI/ML) experts to discuss progress and challenges in the application of AI/ML techniques to current practices in Earth system observation and prediction,” ECMWF’s Massimo Bonavita says. “This is a very active and dynamic field of research, and our goal is to provide the community with a view of cutting-edge work and to encourage collaboration and cross-fertilisation across science domains.”

Thematic focus

The use of ML/deep learning (DL) techniques is becoming widespread in Earth system observation and prediction (ESOP). The scale, complexity and sophistication of ML/DL technologies applied in ESOP has also increased considerably over the last few years.

As a result, ML/DL tools are increasingly integrated into ESOP applications. In some areas, they have the potential to substitute traditional methods.

“Machine learning is, for example, playing a growing role in the use of remotely-sensed observations both in terms of the information that we are able to extract from them, and also in the impact they have when used in our data assimilation systems to initialise model predictions,” says Massimo.

Thematic areas expected to be covered in the workshop include:

1. Machine learning for Earth observations
2. Hybrid machine learning in data assimilation
3. Machine learning for model emulation and model discovery
4. Machine learning for user-oriented Earth science applications
5. Machine learning at the network edge and high-performance computing

You can read up on last year’s workshop on the ECMWF website, and a more detailed report has recently been published by Nature.

Registration and abstracts

The workshop will be held at ECMWF in Reading, UK. While it is intended mainly as an in-person event, remote attendance will be possible.

Registration is open until 14 October, and abstracts can be submitted until 15 August.