RepNN: Tackling spectral bias in deep neural networks for regression and PDE problems via parameter reparameterization
cs.LG, math-ph, math.MP
Submitted: 2026-06-15
Updated: 2026-09-28
Code: https://github.com/UQ4MI/RepNN
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Fourier Neural Operator for Parametric Partial Differential Equations
- Theory of the Frequency Principle for General Deep Neural Networks
- Overcoming Spectral Bias via Cross-Attention
- Enforcing hidden physics in physics-informed neural networks
- On the Activation Function Dependence of the Spectral Bias of Neural Networks
- Frequency-adaptive tensor neural networks for high-dimensional multi-scale problems
- Random Weight Factorization Improves the Training of Continuous Neural Representations
- Multiscale lubrication simulation based on fourier feature networks with trainable frequency
- Adam: A Method for Stochastic Optimization
- Generalized Transferable Neural Networks for Steady-State Partial Differential Equations
- Stable Tensor Neural Networks for Rapid Deep Learning
- General Fourier Feature Physics-Informed Extreme Learning Machine (GFF-PIELM) for High-Frequency PDEs
- Adaptive-Distribution Randomized Neural Networks for PDEs: A Low-Dimensional Distribution-Learning Framework
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks