mlsanityData to 14 Oct 2026

RANDSMAPs: Random-Feature/multi-Scale Neural Decoders with Mass Preservation

arXiv 2601.14794 ↗First posted 21 Jan 2026math.NA cs.LG cs.NAcited by 0

Not on the map. Why?

Abstract (arXiv)

We introduce RANDSMAPs (Random-feature/multi-scale neural decoders with Mass Preservation), numerical analysis-informed, explainable neural decoders designed to explicitly respect conservation laws when solving the challenging ill-posed pre-image problem in manifold learning. We start by proving the equivalence of vanilla random Fourier feature neural networks to Radial Basis Function interpolation and the double Diffusion Maps (based on Geometric Harmonics) decoders in the deterministic limit. We then establish the theoretical foundations for RANDSMAP and introduce its multiscale variant to capture structures across multiple scales. We formulate and derive the closed-form solution of the corresponding constrained optimization problem and prove the mass preservation property. Numerically, we assess the performance of RANDSMAP on three benchmark problems/datasets with mass preservation obtained by the Lighthill-Whitham-Richards traffic flow PDE with shock waves, 2D rotated MRI brain images, and the Hughes crowd dynamics PDEs. We demonstrate that RANDSMAPs yield high reconstruction accuracy at low computational cost and maintain mass conservation at single-machine precision. In its vanilla formulation, the scheme remains applicable to the classical pre-image problem, i.e., when mass-preservation constraints are not imposed.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 24 of its 85 references that are in the map. First posted on arXiv 21 Jan 2026.

Builds on

24 of 85 references.

References · 24

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  2. 2017Deep learning for universal linear embeddings of nonlinear dynamicscited by 494
  3. 2015An Introduction to Matrix Concentration Inequalitiescited by 397
  4. 2015Adversarial Autoencoderscited by 364
  5. 2015On the Equivalence between Kernel Quadrature Rules and Random Feature Expansionscited by 138
  6. 2020Random Features for Kernel Approximation: A Survey on Algorithms, Theory, and Beyondcited by 109
  7. 2020Multiscale Simulations of Complex Systems by Learning their Effective Dynamicscited by 86
  8. 2019Coarse-scale PDEs from fine-scale observations via machine learningcited by 35
  9. 2022Constructing coarse-scale bifurcation diagrams from spatio-temporal observations of microscopic simulations: A parsimonious machine learning approachcited by 25
  10. 2023Sampling weights of deep neural networkscited by 20
  11. 2020LOCA: LOcal Conformal Autoencoder for standardized data coordinatescited by 19
  12. 2022Parsimonious Physics-Informed Random Projection Neural Networks for Initial-Value Problems of ODEs and index-1 DAEscited by 19
12 more
  1. 2022Double Diffusion Maps and their Latent Harmonics for Scientific Computations in Latent Spacecited by 16
  2. 2024RandONet: Shallow-Networks with Random Projections for learning linear and nonlinear operatorscited by 16
  3. 2022Data-driven Control of Agent-based Models: an Equation/Variable-free Machine Learning Approachcited by 15
  4. 2022A Method for Computing Inverse Parametric PDE Problems with Random-Weight Neural Networkscited by 14
  5. 2021On the Parameter Combinations That Matter and on Those That do Notcited by 12
  6. 2024Random Projection Neural Networks of Best Approximation: Convergence theory and practical applicationscited by 11
  7. 2025Enabling Local Neural Operators to perform Equation-Free System-Level Analysiscited by 10
  8. 2024GoRINNs: Godunov-Riemann Informed Neural Networks for Learning Hyperbolic Conservation Lawscited by 8
  9. 2023Nonlinear dimensionality reduction then and now: AIMs for dissipative PDEs in the ML eracited by 7
  10. 2021Randomized Projection Learning Method forDynamic Mode Decompositioncited by 2
  11. 2024On Learning what to Learn: heterogeneous observations of dynamics and establishing (possibly causal) relations among themcited by 2
  12. 2025Next Generation Equation-Free Multiscale Modelling of Crowd Dynamics via Machine Learningcited by 2

Built on by

Not cited yet.