VVZ API is not affiliated with ETH Zurich. Data might be outdated or incorrect. Please view the official ETHZ Vorlesungsverzeichnis for binding information.

651-4918-00L 2 Credits BSC , MSC , GS D-ERDW
You're viewing possible stale or outdated data. Please check the latest semester for more up-to-date information.

Machine Learning for Earth and Planetary Sciences I

Lecturers & Examiners: Dr. Men-Andrin Meier, Dr. Leila Mizrahi
Machine Learning for Earth and Planetary Sciences I focuses on supervised learning and does not require enrollment in Machine Learning for Earth and Planetary Sciences II (unsupervised and probabilistic learning focused) or Machine Learning for Earth and Planetary Sciences III (project focused).
VVZ CR n/a

Last Updated: 2026-02-05 16:38:32

Abstract

This is a 4 week intensive, hands-on course on the basics of machine learning and neural networks. With numerous examples from across the geosciences we learn how to use supervised machine learning techniques to solve geoscience problems in a data driven way.

Objective

The core objective is to understand the logic behind neural network algorithms, and to learn how to use them to tackle your own research problems. This course should prepare you to employ supervised ML techniques e.g. for your BSc or MSc thesis project. You will learn how to use and handle large data sets of various kinds, including point measurements (e.g. ocean temperatures), time series (e.g. seismograms) and image data (e.g. maps). We will go through and develop a wide range of example problems, and learn how to apply different types of neural networks to solve scientific problems.

Content

The class consists of 4 hours of interactive lectures per week, and 4 hours per week of partially supervised exercises. During the lectures we go through, and further develop, a series of jupyter notebooks. In the exercises you apply and modify the concepts learned in class to other, similar problems. Brief introduction to python coding Single- and multi-parameter optimisation ● Grid search ● Linear least squares ● Non-linear least squares Neural networks for regression and classification ● Theory ● Model design ● Model optimisation ● Performance evaluation ● Training and practical considerations Convolutional neural networks Modern architectures overview and application in geosciences ● Recurrent neural networks ● Variational Autoencoders ● Generative and large language models

General Information

Language
English
Levels
BSC , MSC , GS
Frequency
Yearly recurring

Examination

Type
graded semester performance
Graded weekly exercises.

Registration & Places

Max Places
30

Course Components

Type Title Time & Place Hours
lecture with exercise Machine Learning for Earth and Planetary Sciences I
Course takes place during weeks 1 - 4 (last session on Friday 15 March 2024).
  • Wed 14:15-18:00 (NO D 1)
  • Fri 08:15-12:00 (RZ F 21)
32 h semesterly

Offered In