Inferring a posterior over minimal task-sufficient hidden states from one field, then conditioning predictions on it, reduces rollout error in neural PDE simulation by closing 59% of the gap to oracle performance.
Seq.” denotes released trajectories; “seg
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
-
Posterior-First Neural PDE Simulation: Inferring Hidden Problem State from a Single Field
Inferring a posterior over minimal task-sufficient hidden states from one field, then conditioning predictions on it, reduces rollout error in neural PDE simulation by closing 59% of the gap to oracle performance.