ML4EO-workshop-2026

A hands-on workshop on using Terramind, Terrakit and Iterate prepared for the Machine Learning for Earth Observation conference, 2026. There are two notebooks focused on using Geospatial AI Foundation models:

• Introduction to EO Foundation Models: Inference and Fine‑Tuning (1.5 hrs)
• Advanced Geospatial AI: Multimodality, TerraTorch Iterate, and TerraKit (1.5 hrs)

Each notebook can be run in a Jupyter Notebook environment, preferable with access to GPUs.

Prerequisites

1. Python
2. Machine Learning knowledge
3. Some understanding of Geo-spatial datasets.

Getting started

This workshop will make use of Colab as a hosted Jupyter Notebook service. You must be logged in with a Google Account to use the service. Complete the following steps to get started!

1. Navigate to Google CoLab.
2. Select File > Open Notebook
3. Select GitHub and enter https://github.com/IBM/ML4EO-workshop-2026 into the search box. Select one of the two Notebooks to start working.
4. Optionally authorize GitHub to be used with CoLab. This will enable changes to be saved in GitHub but is not necessary to complete the Workshop.
5. Select Additional connection options > Change runtime type > Select T4 GPU > Save.
6. Your ready to go! To run a cell, select it, then either click the grey run triangle or hit ⌘/Ctrl+Enter.

Citations

Sentinel-2. The European Space Agency. 2015. Sentinel-2 Mission. https://sentinel.esa.int/web/sentinel/missions/sentinel2. Accessed: 2026-05-01.

Vertical Aerial Photography. Environment Agency (2025). Available from https://www.data.gov.uk/dataset/4921f8a1-d47e-458b-873b-2a489b1c8165/vertical-aerial-photography

LIDAR Composite Digital Terrain Model. Environment Agency (2025). Available from https://www.data.gov.uk/dataset/e529ca2f-b4ce-403e-8cef-ab821061c4f3/lidar-composite-digital-terrain-model-dtm-2m

The extent and zonation of saltmarsh in England: 2016-2019. Environment Agency, (2022) https://www.gov.uk/government/publications/the-extent-and-zonation-of-saltmarsh-in-england-2016-2019

Attribution

Contains modified Copernicus Sentinel data [2019-2024] © Environment Agency copyright and/or database right 2023. Licensed under the Open Government Licence v3.0.