AD-RaNN learns an effective low-dimensional sampling distribution for hidden parameters in randomized neural networks by optimizing a vector p via PDE-driven or data-driven adaptation and a two-stage least-squares procedure, improving accuracy on benchmark PDE problems.
Title resolution pending
8 Pith papers cite this work. Polarity classification is still indexing.
8
Pith papers citing it
representative citing papers
Parameterizing the temporal derivative in PINNs and reconstructing via Volterra integral yields 100-200x lower errors on advection, Burgers, and Klein-Gordon equations while proving equivalence to the original PDE.
An actor-critic framework built on a time-inhomogeneous little q-function and conditional normalizing flows serves as a mesh-free solver for entropy-regularized jump-diffusion control problems and stochastic games.
Mamba-based neural operators predict stiff chemical kinetics evolution with high fidelity from initial states on Syngas and GRI-Mech 3.0 mechanisms.
PINNs fail on spurious solutions admitted by the residual loss; adaptive pseudo-time stepping with Jacobian-based step selection improves accuracy and robustness on PDE benchmarks.
A block-diagonal symmetrizer and algebraic conditions on closure blocks enable a data-learnable parametrization of ML moment closures for 2D RTE that guarantees symmetrizable hyperbolicity by construction.
ResearchEVO automates the discover-then-explain cycle by evolving algorithms via fitness-driven LLM co-evolution and generating grounded, anti-hallucination research papers through sentence-level RAG.
A single-layer architecture called FlowMixer uses constrained matrix operations and a semi-group property to enable depth-agnostic, interpretable spatiotemporal forecasting with direct eigenmode extraction.
citing papers explorer
-
Adaptive-Distribution Randomized Neural Networks for PDEs: A Low-Dimensional Distribution-Learning Framework
AD-RaNN learns an effective low-dimensional sampling distribution for hidden parameters in randomized neural networks by optimizing a vector p via PDE-driven or data-driven adaptation and a two-stage least-squares procedure, improving accuracy on benchmark PDE problems.
-
Learning on the Temporal Tangent Bundle for Physics-Informed Neural Networks
Parameterizing the temporal derivative in PINNs and reconstructing via Volterra integral yields 100-200x lower errors on advection, Burgers, and Klein-Gordon equations while proving equivalence to the original PDE.
-
An Actor-Critic Framework for Continuous-Time Jump-Diffusion Controls with Normalizing Flows
An actor-critic framework built on a time-inhomogeneous little q-function and conditional normalizing flows serves as a mesh-free solver for entropy-regularized jump-diffusion control problems and stochastic games.
-
Kinetic-Mamba: Mamba-Assisted Predictions of Stiff Chemical Kinetics
Mamba-based neural operators predict stiff chemical kinetics evolution with high fidelity from initial states on Syngas and GRI-Mech 3.0 mechanisms.
-
When PINNs Go Wrong: Pseudo-Time Stepping Against Spurious Solutions
PINNs fail on spurious solutions admitted by the residual loss; adaptive pseudo-time stepping with Jacobian-based step selection improves accuracy and robustness on PDE benchmarks.
-
Machine learning moment closure models for the radiative transfer equation IV: enforcing symmetrizable hyperbolicity in two dimensions
A block-diagonal symmetrizer and algebraic conditions on closure blocks enable a data-learnable parametrization of ML moment closures for 2D RTE that guarantees symmetrizable hyperbolicity by construction.
-
ResearchEVO: An End-to-End Framework for Automated Scientific Discovery and Documentation
ResearchEVO automates the discover-then-explain cycle by evolving algorithms via fitness-driven LLM co-evolution and generating grounded, anti-hallucination research papers through sentence-level RAG.
-
FlowMixer: A Depth-Agnostic Neural Architecture for Interpretable Spatiotemporal Forecasting
A single-layer architecture called FlowMixer uses constrained matrix operations and a semi-group property to enable depth-agnostic, interpretable spatiotemporal forecasting with direct eigenmode extraction.