A framework that automatically selects manifold learning methods and hyperparameters on subgraphs recovers reduced-order dynamics from spatial-temporal PDE data faster and often more accurately than manual tuning.
Discovering governing equations from partial measurements with deep delay autoencoders
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Automated Manifold Learning for Reduced Order Modeling
A framework that automatically selects manifold learning methods and hyperparameters on subgraphs recovers reduced-order dynamics from spatial-temporal PDE data faster and often more accurately than manual tuning.