GeoMAE applies masking-based self-supervised learning to spatio-temporal graphs to forecast under missing values and reports up to 13.2% gains over baselines on real datasets.
Neural odes for informative missingness in multivariate time series,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
GeoMAE: Masking Representation Learning for Spatio-Temporal Graph Forecasting with Missing Values
GeoMAE applies masking-based self-supervised learning to spatio-temporal graphs to forecast under missing values and reports up to 13.2% gains over baselines on real datasets.