$$
\varsigma~Trending~Now~\jmath
$$
Multi-resolution reconstruction of longitudinal streambed footprints using embedded sparse convolutional autoencoders
https://www.sciencedirect.com/science/article/pii/S0022169425001908?via%3Dihub
<aside> <img src="/icons/duck_yellow.svg" alt="/icons/duck_yellow.svg" width="40px" /> Keypoints:
✅ AI-driven multi-resolution reconstruction
✅ Reliable under sparse and corrupted data
✅ A step towards digital twinning of river systems
We introduce a convolutional autoencoder framework to reconstruct high-resolution streambed footprints from sparse data. Our model leverages deep learning, transfer learning, and attention mechanisms to recover fine-scale riverbed patterns from sparse observations—even under severe data loss.
</aside>
Digital twinning of river basins towards full-scale, sustainable and equitable water management and disaster mitigation
https://www.nature.com/articles/s44304-024-00047-2
<aside> <img src="/icons/duck_yellow.svg" alt="/icons/duck_yellow.svg" width="40px" /> Keypoints: