Communication Dans Un Congrès Année : 2025

A hackathon for flood map prediction from geospatial data with parsimonious machine learning models

Résumé

Flooding poses significant risks across various sectors in France. This paper presents the outcomes of a machine learning hackathon focused on predicting the extent of various types of floods by leveraging a combination of geospatial and climate data. A Convolutional Neural Network (CNN) emerged as the most effective model, achieving strong performance in predicting the temporal evolution of flood risk maps. The evaluation not only includes prediction accuracy but also incorporates robustness, frugality, and explainability, in line with the principles of trustworthy AI principles. A key feature of this challenge was the absence of streamflow data, allowing the models to predict floods in regions where such data is unavailable. This highlights the potential of machine learning to improve flood forecasting in data-scarce environments.

2 Hackathon setup 2.1 Geospatial Data Labels: Flood maps were extracted from the Sen1Floods11 dataset [2] which can be visualized on the Global Flood Database. Sen1Floods11 provides surface water data, including raw Sentinel-1 WIP paper.

Fichier principal
Vignette du fichier
ScientificDesign_Hackathon_Flooding (1).pdf (1.13 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-04813418 , version 1 (11-12-2024)

Licence

Identifiants

  • HAL Id : hal-04813418 , version 1

Citer

David Medernach, Cyril Lemaire, Eva Girousse, Julie Keisler, Julie Richon, et al.. A hackathon for flood map prediction from geospatial data with parsimonious machine learning models. ICLR 2025 Workshop on Tackling Climate Change with Machine Learning, Climate Change AI, Apr 2025, Singapore, France. ⟨hal-04813418⟩
224 Consultations
334 Téléchargements

Partager

  • More