A weak-form regression framework using spatial Gaussian kernels removes bias in recovering drift b(x) and diffusion a(x) for stochastic generators from single sparse regressions, validated on benchmarks with low coefficient and density errors.
Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Weak-Form Recovery of Stochastic Generators and Dynamical Invariants
A weak-form regression framework using spatial Gaussian kernels removes bias in recovering drift b(x) and diffusion a(x) for stochastic generators from single sparse regressions, validated on benchmarks with low coefficient and density errors.