Best practices in deep learning for oceanography
Co-organized by ESSI6/OS4
Convener:
Julien Brajard
|
Co-conveners:
Aida Alvera-Azcárate,
Alexander Barth,
Rachel Furner,
Matjaz Licer
Fri, 08 May, 10:45–12:30 (CEST) Room -2.82
Fri, 10:45
In this short course, we aim to present a set of best practices for applying and assessing deep learning methods in oceanographic research. We will also highlight common pitfalls and how to avoid them.
The course will be structured around a series of short presentations and practical examples covering key topics, including:
- Types of oceanographic problems suited for deep learning: reconstruction, prediction, …
- Building datasets appropriate for deep learning applications: constitution of - training/validation/test datasets, effect of non-stationnarity, type/quality/number of data, …
- Training strategies and model selection: normalization, supervised training, generative models, …
- Validation and evaluation of ocean products derived from deep learning: accuracy, realism, …
- Ethical considerations: reproducibility, open science, and the environmental impact of deep learning
The course welcomes participants with varying levels of experience, from first‑time users to those already applying deep learning in their work. We adopt a broad, practical approach, highlighting best practices that are transferable across oceanographic problems.
Hands‑on exercises will be based on open notebooks available at
https://github.com/RachelFurner/ToySSTProblem
A cloud computing solution will be provided so that participants can run the notebooks without requiring a local installation. Please bring your laptop!
Session assets
Speakers
- Matjaz Licer, Slovenian Environment Agency, Slovenia
- Redouane Lguensat, Institut Pierre-Simon Laplace, France
- Anass El Aouni