Physics-Informed Neural Plasticity: PDE Solvers That Reshape Themselves
Its authors withdrew this paper, so there is no full text to read or summarize.
Lists
Only you see this
Abstract
Physics-informed neural PDE solvers adapt their parameters to satisfy governing equations, yet their \emph{representational structure typically remains fixed throughout training}. This rigidity is poorly matched to PDE solutions with strongly heterogeneous complexity across space and space--time, leaving capacity insufficient where the physics is difficult and redundant where it is simple. We introduce \emph{physics-informed neural plasticity}, a paradigm in which the representation itself reshapes during optimization in response to unresolved physics. We instantiate this principle with \emph{Representation Capacity Adaptation for PDEs} (ReCAP), a Gaussian-localized solver that dynamically redistributes capacity through local enrichment, residual-directed splitting, gate-based pruning, and function-aware merging. ReCAP uses responsibility-weighted error indicators and the geometry of residual energy to determine where and how to refine. To limit the disturbance introduced by splitting, we introduce \emph{quiet-child refinement}, which initializes new components by transporting the parent representation while controlling instantaneous functional perturbation. We further establish conditional a posteriori reliability and structural-stability guarantees linking localized physics residuals to solution error and stable refinement. Across five challenging 3D and 4D PDE benchmarks against 11 physics-informed solvers, ReCAP achieves the lowest relative $L^2$ error on every problem, reducing error by $10.7\%$--$27.5\%$ relative to the strongest competing result. These results suggest that physics-informed solvers need not merely learn their parameters---they can learn how their representational capacity should be organized.
BibTeX
@misc{cheng2027physics,
author = {Chun-Wun Cheng and Bingcheng Hu and Angelica I Aviles-Rivero},
title = {{Physics-Informed Neural Plasticity: PDE Solvers That Reshape Themselves}},
year = {2027},
url = {https://openreview.net/forum?id=e54TpkBQ8D},
howpublished = {OpenReview},
note = {Submitted to ICLR 2027},
}Loading related content...