Introduction to Machine Learning for Climate Scientists

4 Mar 2024, 13:00 → 5 Mar 2024, 16:00 Europe/Berlin

Room 034 (DKRZ Main Building)

Adeniyi Mosaku (DKRZ / HEREON), Etienne Plésiat, Harsh Grover (DKRZ / HEREON), Johannes Meuer (DKRZ), Paul Keil (DKRZ / HEREON)

Target group: beginner in ML with some Python experience

Deadline for registration: January 31 15:00 CET

Machine learning is becoming a popular method for climate scientists. While there are many tutorials and courses available, researchers often face challenges when applying the tutorial concepts to actual climate data, that can be quite different from standard machine learning datasets.

Therefore, we are offering an Introduction to Machine Learning for Climate Scientists at DKRZ. The course will be held in person, and the number of participants is limited to 20. 

The course will be held from March 4 13:00 to March 5 16:00 at DKRZ seminar room 034. Coffee breaks will be provided during the workshop.

Participants are required to have a working knowledge of Python.

Architectures and Applications.pdf

hands-on_pytorch.zip

Introduction.pdf

setup_pytorch_2.pdf

XAI.zip

Monday, 4 March

Day 1

Conveners: Adeniyi Mosaku (DKRZ / HEREON), Harsh Grover (DKRZ / HEREON), Paul Keil (DKRZ / HEREON)

1 Introduction to Machine Learning

• A comprehensive overview of the concepts and principle behind Machine Learning
• Exploration of real-world applications of Machine Learning
• Differentiating different Machine Learning types
• Introducing popular Machine Learning tools and frameworks

Speaker: Adeniyi Mosaku (DKRZ)

14:40 Coffee Break

2 Architectures and Applications

• An overview of state of the art Machine Learning Methods
• Examples from weather, climate and beyond

Speaker: Paul Keil

3 Explainable AI

• Introduction to Explainable AI
• Importance of Explainability
• Interpretability techniques and use cases

Speaker: Harsh Grover

Tuesday, 5 March

Day 2

Conveners: Etienne Plésiat, Johannes Meuer (DKRZ)

4 PyTorch applied to Climate Science

• Setup of the accounts
• Introduction to PyTorch with examples
• Definition of the task
• Creation of the training, validation and test datasets

Speaker: Etienne Plésiat

10:30 Coffee Break

5 PyTorch applied to Climate Science

• Building the CNN
• Training the model
• Testing the model

Speaker: Etienne Plésiat

12:00 Lunch break

6 Advanced ML use-case: Reconstructing missing climate data

• Create and modify inpainting CNN for reconstructing climate data
• Train the model with different configurations
• Validate the model on test data

Speaker: Johannes Meuer (DKRZ)

7 Closing remarks