A contrastive learning method learns stable representations from rainfall and terrain data to improve short-term landslide risk estimates under rainfall forecast displacement, reporting up to 37% higher precision than baselines on Japanese data.
Pyramidal implementation of the affine lucas kanade feature tracker description of the algorithm,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Learning Displacement-Robust Representations for Landslide Early Warning under Rainfall Forecast Uncertainty
A contrastive learning method learns stable representations from rainfall and terrain data to improve short-term landslide risk estimates under rainfall forecast displacement, reporting up to 37% higher precision than baselines on Japanese data.