Selected work

Representative Papers

Non-stationary GEV models for estimating design sea-states in a changing climate. Applications to offshore wind farms along the French coasts

Mar 09, 2026

This study addresses the inadequacy of conventional extreme-event design methods—based on stationarity assumptions—in ensuring the durability of offshore wind structures in France under nonstationary sea conditions driven by climate change. To overcome this limitation, a nonstationary generalized extreme value (GEV) model is developed, integrating CMIP6 multi-model ensemble projections with reanalysis data to quantify future changes in extreme sea states using monthly maxima of significant wave height. The work further introduces a lifetime-equivalent design sea state framework for forward-looking structural design. Results indicate intensified winter extremes and attenuated summer extremes in the Atlantic and English Channel, leading to heightened seasonal contrasts, while trends in the Mediterranean remain uncertain. Overall, design conditions are projected to become more severe, underscoring the necessity of abandoning stationarity assumptions to enhance climate resilience in offshore wind infrastructure.

0 citationsRead paper

Efficient Self-Supervised Learning for Earth Observation via Dynamic Dataset Curation

Apr 09, 2025

To address pervasive redundancy and long-tailed class distributions in remote sensing imagery, this paper proposes a dynamic dataset pruning framework for self-supervised learning (SSL) on SAR data—requiring no pre-trained feature extractor. The method performs online diversity assessment and iterative sample reweighting to dynamically optimize the composition of the full decade-long Sentinel-1 WV archive, simultaneously ensuring intra-class and inter-class balance while enhancing representation robustness. It achieves the first end-to-end SSL pretraining specifically for SAR imagery and releases Nereus-SAR-1—the first foundation model tailored for marine SAR analysis. Evaluated on three downstream tasks, Nereus-SAR-1 improves linear probe accuracy by an average of 5.2% and accelerates training efficiency by 37%. All model weights are publicly released.

0 citationsRead paper

Recent publications

Latest Papers

Non-stationary GEV models for estimating design sea-states in a changing climate. Applications to offshore wind farms along the French coasts

Mar 09, 2026

This study addresses the inadequacy of conventional extreme-event design methods—based on stationarity assumptions—in ensuring the durability of offshore wind structures in France under nonstationary sea conditions driven by climate change. To overcome this limitation, a nonstationary generalized extreme value (GEV) model is developed, integrating CMIP6 multi-model ensemble projections with reanalysis data to quantify future changes in extreme sea states using monthly maxima of significant wave height. The work further introduces a lifetime-equivalent design sea state framework for forward-looking structural design. Results indicate intensified winter extremes and attenuated summer extremes in the Atlantic and English Channel, leading to heightened seasonal contrasts, while trends in the Mediterranean remain uncertain. Overall, design conditions are projected to become more severe, underscoring the necessity of abandoning stationarity assumptions to enhance climate resilience in offshore wind infrastructure.

0 citationsRead paper

Efficient Self-Supervised Learning for Earth Observation via Dynamic Dataset Curation

Apr 09, 2025

To address pervasive redundancy and long-tailed class distributions in remote sensing imagery, this paper proposes a dynamic dataset pruning framework for self-supervised learning (SSL) on SAR data—requiring no pre-trained feature extractor. The method performs online diversity assessment and iterative sample reweighting to dynamically optimize the composition of the full decade-long Sentinel-1 WV archive, simultaneously ensuring intra-class and inter-class balance while enhancing representation robustness. It achieves the first end-to-end SSL pretraining specifically for SAR imagery and releases Nereus-SAR-1—the first foundation model tailored for marine SAR analysis. Evaluated on three downstream tasks, Nereus-SAR-1 improves linear probe accuracy by an average of 5.2% and accelerates training efficiency by 37%. All model weights are publicly released.

0 citationsRead paper