IAMG2026
The 24. annual conference of the IAMG
August 23 - 28, 2026, Montreal, Canada

Short Courses

Information about registration to the short courses coming soon.

Mathematical Morphological (Spatial) Algorithms in Surfaces↑

  • organized by: B. S. Daya Sagar
  • Venue: coming soon
  • Date: August 23, 2026
  • Time: coming soon
  • Min. number of participants: 5
  • Max. number of participants: 15
  • costs per participant: coming soon
  • Attendance Requirements: -
  • Recommended prerequisites:

Introduction

Many micro-scale to macro-scale surfaces represented in numerous ways—(e.g.: Digital Elevation Models: an Important Source of Data for Geoscientists, IEEE Geoscience and Remote Sensing Magazine, v. 8, no. 4, p. 138-142. 10.1109/MGRS.2020.3031910; rock surfaces; floodplain surfaces, material and mineral surfaces, etc))—contains rich geometric, morphologic and topologic information, but hidden for naked eyes. Unraveling such information from the surfaces of relevance to the Earth and Planetary sciences is a challenge. Such information from time-dependent and time independent surfaces is essential to model and visualize terrestrial and planetary surficial phenomena and processes. Geoscientists with appropriate mathematical knowledge can better exploit the full potential of the surfaces that have hitherto been analyzed via classical mathematical and statistical techniques. Mathematical Morphology is an area of geoscience that most people don’t realize will literally change the way they look at the surfaces of relevance to Earth and Planetary studies! This short course would provide a glimpse of how mathematical morphology could be employed to treat surfaces to derive scientific outcomes. Geoscience communities involving all its sub-branches some way or the other are familiar with surfaces of varied types acquired through varied sensing mechanisms. Mathematical morphology, a powerful artificial intelligence tool, is the basis for developing the spatial morphological algorithms that deal with those surfaces of relevance to the Earth and Planetary sciences. With the advent of the availability of accurate data at multiple spatial, spectral and temporal scales from various sources, the simulation, reconstruction, and prediction of surficial changes and processes on discrete time scales are feasible. With over three and a half decades of knowledge in showing applications of mathematical morphology, fractal geometry and chaos theory in understanding terrestrial phenomena and processes, Professor Daya Sagar deals with first lecturing on Mathematical Morphology, followed by analyzing such surfaces through mathematical morphology. This proposed tutorial would provide required knowledge to the interested audience on mathematical morphology in analyzing surfaces in a firm quantitative manner.
  • Audience would gain knowledge on morphological thinking to better understand the terrestrial and planetary surfaces processes.
  • Audience would be able to employ mathematical morphological operations and transformations in studies related to geosciences, remote sensing and geospatial data sciences.
  • Audience would be able to reason the importance of differential morphology in the context of studying the surficial morphological changes of relevance to terrestrial and planetary studies, with examples on rock surfaces, porous-medium, DEMs.
  • Audience with reasonable knowledge in basic set theory and the surfaces would benefit significantly in understanding the morphological equations, illustrations, and spatial (morphological) algorithms that would be explained in a span of six hours.

Machine learning tools for mineral systems modelling and mineral predictive mapping↑

  • organized by: Geological Survey of Finland (GTK), Finland and Beak Consultants GmbH, Germany
  • Venue: coming soon
  • Date: August 23, 2026
  • Time: coming soon
  • Min. number of participants: 5
  • Max. number of participants: 15
  • costs per participant: coming soon
  • Attendance Requirements: -
  • Recommended prerequisites:

Topic introduction:

Machine learning has proven to be applied successfully to mineral exploration targeting in the last decade. The aim of this short course is to provide an introduction into artificial intelligence applied to mineral predictive mapping to generate drilling targets during mineral exploration and to estimate mineral resource for the assessment of critical raw materials.

Course description:

In the first part of the short course, a short review will be provided about the concept of mineral predictive mapping. Following up, different machine learning algorithms are being presented and discussed on how they may be used for mineral predictive mapping. The second main part of the short course is strongly focusing on the practical hands-on-training for a newly developed open source tool for mineral predictive mapping (MPM) for exploration and assessment of critical raw materials called “EIS” – Exploration Information System. The demonstrated software tool is open-source and will be able to run for free on the computer of the participants.

Specific topics of the short course include:

  • Short introduction to mineral predictive mapping concepts,
  • Review of machine learning algorithms applied to mineral predictive mapping, e.g. neural networks, self-organizing maps, fuzzy logic, random forest,
  • Hands-on-training on open source tool “EIS” for mineral systems approach and mineral predictive mapping:
    “EIS” provides a library of tools for using mineral systems approach and mineral predictive mapping in a combined graphical user interface. It runs under Quantum GIS (QGIS) on Windows systems. Training tutorial starts from installation of EIS Toolkit and EIS Plugin on the computer leading up to the final compilation of a mineral prospectivity map using the EIS Wizard in QuantumGIS,
  • Case studies for mineral predictive mapping from Europe and Northern America
  • Presentation of on-going new developments in mineral exploration and prospectivity mapping.

Top “takeaways”:

This short course will provide best practices, new ideas and insights in mineral exploration, especially with regards to application of the newly developed open source for mineral predictive mapping. At the end of the course, participants will be able to run mineral predictive mapping on their own computer using the provided and trained open-source software tools.

Practical use of multiple-point statistics algorithms - within Python - to generate heterogeneous 3D property fields↑

  • organized by: Guillaume Pirot, School of Earth and Oceans, Centre for Exploration Targeting, The University of Western Australia, Australia
  • Venue: coming soon
  • Date: August 23, 2026
  • Time: coming soon
  • Min. number of participants: 5
  • Max. number of participants: 15
  • costs per participant: coming soon
  • Attendance Requirements: Participants are required to bring their own laptop. Prior to the workshop, participants should have a functioning Anaconda distribution https://www.anaconda.com/download installed on their laptop.
  • Recommended prerequisites: Participants should have some basic knowledge of coding with Python, that they can get by following the fundamentals of Python tutorials available at https://swcarpentry.github.io/python-novice-inflammation/index.html

Objectives

Spatial heterogeneity is a key driver of groundwater flow and transport. Despite the existence of multiple tools to generate geologically realistic property fields, those are not sufficiently used. The main objective of this workshop is to familiarise practitioners with the use of multiple-point statistics algorithms that enable the generation of complex spatial pattern from a training image while honouring local observations.

Content

Introduction to multiple-point statistics (MPS), familiarization with the geone Python library, bring your own project to work on or modelling of the Lower Burdekin Delta aquifer using MPS conditionally to interpretation of legacy boreholes.

Fundamental Deep Learning Concepts for Applied Geoscientists↑

  • organized by: Xiao Xia Liang, Department of Civil and Environmental Engineering, Princeton University, Princeton, NJ, United States and Dany Lauzon, Tao Wen
  • Venue: coming soon
  • Date: August 23, 2026
  • Time: coming soon
  • Min. number of participants: 5
  • Max. number of participants: 15
  • costs per participant: coming soon
  • Attendance Requirements: -
  • Recommended prerequisites:

This short course focuses on the fundamental concepts of deep learning (DL). Topics include data cleaning and preparation, activation functions, hyperparameter tuning, model optimization through forward and backward propagation, and regularization techniques. These core concepts will be explored through DL models applied to geoscience case studies. Students will gain hands-on experience implementing these methods and develop a deeper understanding of how model design choices influence prediction quality and uncertainty.

Several DL architectures relevant to geoscience applications will be introduced. Students will begin with Multilayer Perceptrons (MLPs) to establish the theoretical foundations of deep learning, followed by a Convolutional Neural Network (CNN) to demonstrate spatial feature extraction from geological data. The course will then cover a Recurrent Neural Network (RNN) Autoencoder and a Variational Autoencoder (VAE) to illustrate representation learning, latent space interpretation, and probabilistic modeling of subsurface uncertainty.

The objective of this course is to teach the fundamentals of deep learning to anyone who wishes to enter into the field or deepen their understanding of its core concepts through the development and application of multiple DL architectures.

Machine Learning for Geostatisticians↑

  • organized by: Michael J. Pyrcz, The University of Texas at Austin
  • Venue: coming soon
  • Date: August 23, 2026
  • Time: coming soon
  • Min. number of participants: 5
  • Max. number of participants: 15
  • costs per participant: coming soon
  • Attendance Requirements: -
  • Recommended prerequisites:

Description

Machine learning is emerging as a powerful tool for enhancing geostatistical workflows. This is not a departure from tradition—geostatistics has always evolved by integrating theory and methods from other disciplines. If a method helps solve the problem, then it belongs in geostatistics. Spend a day with Professor Michael Pyrcz of The University of Texas at Austin as we build the theory and practice of machine learning, progressing from probability and data analytics to inferential and predictive modeling. Throughout the course, we will explicitly connect machine learning concepts to geostatistical principles, culminating in a discussion of how geostatistics can, in turn, improve machine learning workflows. Learning will be maximized through comprehensive notes, well-documented demonstrations, and interactive dashboards designed to make abstract concepts concrete and actionable.

Public Geoscience Data and Machine Learning Applications↑

  • organized by: Geological Survey of Canada (Mohammad Parsa, Steven E. Zhang, Julie E. Bourdeau, Dianne Mitchinson) and Department of Earth Sciences, Carleton University, Canada (Christopher J.M. Lawley)
  • Venue: coming soon
  • Date: August 23, 2026
  • Time: coming soon
  • Min. number of participants: 5
  • Max. number of participants: 15
  • costs per participant: coming soon
  • Attendance Requirements: -
  • Recommended prerequisites:

Introduction

The short course “Public Geoscience Data and Machine Learning Applications” will provide participants with a practical and accessible entry point into modern, data-driven approaches to mineral exploration. The course begins with an overview of publicly available geoscience datasets, explaining how they are collected and why they are valuable for research and industry. Participants will then learn how to access and obtain large geological, geophysical, and geochemical datasets that are freely available for exploration and scientific analysis. The presenters will introduce the central research topic of this course: mineral prospectivity mapping, a methodology aimed at predicting areas with high potential for mineral deposits. Through real examples, attendees will gain insights into how geoscientists integrate diverse datasets, develop conceptual models, and use data-driven methods to support exploration decisions. The course will then explore two key data domains widely used in prospectivity analysis: geophysical and geochemical data. Participants will learn how these datasets contribute to understanding geological processes that are relevant to mineralization.

A major component of the course is hands-on training. Using the free, easy to use, and visual data-mining and machine learning platform Orange (https://orangedatamining.com), participants will import, preprocess, and analyze geophysical and geochemical datasets. Step-by-step instructions will guide them through building workflows for data cleaning, feature extraction, visualization and exploratory analysis. The course further introduces essential concepts in geodata science and machine learning, emphasizing how these tools can be applied to mineral prospectivity mapping. Participants will implement machine learning algorithms within Orange to create predictive models, evaluate their performance, and generate prospectivity maps based on real-world datasets. By the end of the course, attendees will have gained both theoretical understanding and practical skills that enable them to work confidently with public geoscience data and apply machine learning techniques to mineral exploration challenges.

This short course is designed for anyone interested in applying modern data-driven methods to geoscience and mineral exploration. It is especially suitable for:

  • Geoscientists and exploration geologists who want to learn how to integrate public geoscience datasets with machine learning techniques for mineral prospectivity mapping.
  • Geophysicists and geochemists seeking to expand their skills into data processing, analysis, and interpretation using accessible, user-friendly tools.
  • Graduate students and early-career researchers in geology, geophysics, geochemistry, or data science who want practical, hands-on experience with real-world geoscience data.
  • Mining and exploration professionals looking to incorporate machine learning and geodata science into exploration workflows. No advanced programming background is required. The course uses Orange (https://orangedatamining.com), a free, easy to use, visual data-mining platform, making the course accessible to participants with varying levels of technical experience.

Contents

  • Introduction to Public Geoscience Data Course
  • Introduction to the Software package platform, Orange
  • Public Geophysical Data and Data Processing Applications
  • Public Geochemical Data and Data Processing Applications
  • GeoData Science and Mineral Prospectivity Mapping

Novel constrained potential field data inversion techniques applied to exploration↑

  • organized by: Jeremie Giraud
  • Venue: coming soon
  • Date: August 23, 2026
  • Time: coming soon
  • Min. number of participants: 5
  • Max. number of participants: 15
  • costs per participant: coming soon
  • Attendance Requirements: bring our own laptop
  • Recommended prerequisites: The course is intended for geoscientists with experience in the modelling and interpretation of geophysical data who are interested in testing new tools and in learning about and gaining exposure to recent developments. Participants should have some knowledge of geophysical inversion and be familiar with the basics of Python.

Gaps filled by the course

It fills a gap as no course proposes the teaching of similar methods and training on the usage of similar open source codes, which are applicable to both mineral exploration, oil and gas, and environmental contexts. Methods not elsewhere covered include:

  1. application of petrophysical constraints to the inversion of petrophysical data
  2. exploration of alternative models while maintaining data fit (null-space exploration)
  3. 3D geometrical inversion and trans-dimensional inversion
  4. teaching using open source codes
Existing tools and methods will be made more accessible to practitioners. The presented methods and codes correspond to recent development aimed at filling a capability gap in terms of our capability to:
  1. Integrate various types of information in the inversion of potential data quantitatively;
  2. Quantitatively assess uncertainty, generate alternative scenarios fitting the data, and estimate the range of unknown unknowns.

Purpose and Material

This short course aims to provide participants with some theoretical knowledge and hands-on experience in some of the latest advances in 3D gravity and magnetic inversion techniques. It is designed for geoscientists familiar with inversion to learn recently developed methods that are currently being deployed with partners from the industry and academia. By leveraging tools developed through the MinEx CRC, Loop 3D, and in collaboration with several partner projects, the course is aimed at equipping users with the theoretical background and computational frameworks needed to invert potential field data for exploration and geological applications. In particular, the emphasis is placed on:

  1. the integration with other datasets including petrophysical measurements, geological models, and other geophysical techniques;
  2. introduction to geometrical and trans-dimensional inversion;
  3. consideration of uncertainty and the generation of many models fitting the data equally well.

Course Description

This short course introduces geoscientists to recent methods for potential field inversion developed to mitigate the limitations in the currently available commercial codes, starting with physical property inversion (i.e., inverting for density or magnetic susceptibility), followed by geometrical inversion (i.e., inverting for the geometry of rock units). This course provides an overview of unconstrained inversion techniques and of the application of advanced petrophysical constraints for physical property inversion using the Tomofast-x inversion engine. Geometrical inversion approaches such as level set and trans-dimensional inversion are also covered. Techniques are presented that allow running large scale inversion with, e.g., wavelet compression or efficient sub-sampling of the data to invert.To complement the presented inversion approaches, the course also covers concept of null space - the set of model variations that leave the data misfit unchanged or nearly unchanged - and how this can be leveraged to generate alternative models. Methods to use null space navigation as a tool for model exploration, scenario generation and uncertainty assessment are presented and hands-on exercises guide participants through theoretical concepts and field case studies. Practical examples of recently developed codes rely on the Python language and run in the cloud using Google Colab and Jupyter Notebooks. The codes used for the practicals are open source, enabling participants to continue developing and applying these techniques independently after the course.

Course Outline

09:00-09:20 TALK Introduction
Overview of current state of inversion research globally, brief review of available codes, and specifics of what will be covered.
09:20-09:45 TALK
Introduction to unconstrained inversion and ADMM constraints using Tomofast-x. General presentation of the Tomofast-x open-source potential fields inversion code, and introduction to the ADMM petrophysical bound constraints with field application.
09:45-10:15 HANDS ON
Tomofast-x inversion & ADMM (part 1)
10:15-10:30 Morning tea
10:30-11:00 HANDS ON
Tomofast-x unconstrained inversion & ADMM (part 2)
11:00-11:30 TALK
Introduction to null space analysis, the example of gravity and magnetics Exploring the concept of "null space", how to perturb a model without changing (too much) its misfit and generate new solutions quickly.
11:30-12:00 TALK
Introduction to geometrical inversions using level-sets, presentation of a field example
12:00-13:00 Lunch
13:00-13:30 TALK
Case study in the Pyrenees using gravity analysis of slab subduction Field application of null space navigation to investigate several geological scenarios.
13:30-14:00 HANDS ON
Null space navigation (part 1)
Synthetic models using gravity and magnetic data -- density and magnetic susceptibility perturbations.
14:00-14:45 HANDS ON
Null space navigation (part 2) Field application using gravity data -- geometry perturbation and model sampling.
14:45-15:00 Afternoon tea
15:00-15:15 HANDS ON
catch up on either null-space navigation or Tomofast-x inversion
15:15-15:30 TALK+DEMO
efficient, statistics-preserving data subsampling
15:30-16:00 TALK
Introduction to trans-dimensional inversion
15:00-16:15 DEMO
trans-dimensional inversion and analysis of results

Recommended Reading

  1. Ogarko, V., Frankcombe, K., Liu, T., Giraud, J., Martin, R., and Jessell, M.: Tomofast-x 2.0: an open-source parallel code for inversion of potential field data with topography using wavelet compression, Geosci. Model Dev., 17, 2325–2345, https://doi.org/10.5194/gmd- 17-2325-2024, 2024.
  2. Giraud, J., Ford, M., Caumon, G., Ogarko, V., Grose, L., Martin, M., Cupillard, P.: Geologically constrained geometry inversion and model space navigation to explore alternative geological scenarios: a case study in the Western Pyrenees, Geophysical Journal International, Volume 239, Issue 3, December 2024, Pages 1359–1379, 2024.
  3. Ogarko, V., Giraud, J., Martin, R., and Jessell, M.: Disjoint interval bound constraints using the alternating direction method of multipliers for geologically constrained inversion: Application to gravity data, GEOPHYSICS 86: G1-G11, https://doi.org/10.1190/geo2019-0633.1, 2021.
  4. Giraud, J., Rashidifard, M., Ogarko, V., Pirot, G., Portes dos Santos, L., Caumon, G., Grose, L., Herrero, J., Cupillard, P., Lindsay, M., Jessell, Mark; Aillères, L.: Transdimensional geometrical inversion: Application to...

Multifractals in geophysics and geology↑

  • organized by: Shaun Lovejoy, Physics, McGill University, Canada and Qiuming Cheng, Sun Yat-Sen University Zhuhai, China
  • Venue: coming soon
  • Venue: coming soon
  • Date: August 23, 2026, half day course
  • Time: coming soon
  • Min. number of participants: 5
  • Max. number of participants: 15
  • costs per participant: coming soon
  • Attendance Requirements: -
  • Recommended prerequisites:

The distribution of minerals, of seismic events, the movement of tectonic plates, the Earth's topography, the evolution of biota - and much else - are the product of nonlinear dynamical systems operating over wide ranges of scale in space and in time. These processes are typically scale invariant involving geometric fractal sets of points and – more generally - multifractal fields, measures.
In this three hour course, we survey:

  • The basic statistical properties of multifractal processes, in particular, their exponents that characterize the statistical moments and probability distributions.
  • We discuss anisotropic scaling needed for realistic textures, geomorphologies.
  • Extreme events (scaling in probability space).
  • Numerical modelling (including fractional equations).
  • How dynamical regimes are objectively defined by scaling (particularly the bio-geo megaclimate regime).
  • Data analysis techniques, focusing on Haar fluctuation analysis.
We give examples including plate tectonics, the climate, multifractal geochronologies and macro evolution.