mlsanityData to 14 Oct 2026

Nonlocal Hamiltonian Dynamics on Sparse L'evy Graphs: Spectral Analysis and Multimodal Sampling

arXiv 2610.06904 ↗First posted 1 Oct 2026math.DS cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

We develop a sparse graph method for transporting probability mass toward multimodal target distributions through damped nonlocal Hamiltonian dynamics. The formulation combines logarithmic-mean mobility with symmetric L'evy-type interaction weights, coupling the evolving density to an edge momentum field. A graph constructed from nearest-neighbor connections and sampled long-range edges provides direct mass exchange between spatially separated regions. Once the graph is constructed, the density evolution is deterministic, and each update costs linear in the number of nodes and the long-range sampling budget. Linearization around the target distribution yields a damped oscillator governed by a weighted graph Laplacian. Its spectrum characterizes the interaction between nonlocal connectivity and inertia, with the L'evy exponent alpha tuning the nonlocal connectivity: the spectral gap determines the optimal asymptotic damping, while the largest eigenvalue governs the time-step stability. Experiments on synthetic multimodal distributions demonstrate improved mode balance and more stable mode coverage relative to first-order and MCMC baselines. The resulting framework provides a sparse implementation of nonlocal inertial density transport for sampling problems with low-dimensional spatial structure.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 19 of its 39 references that are in the map. First posted on arXiv 1 Oct 2026.

Builds on

19 of 39 references.

References · 19

  1. 2017Towards Principled Methods for Training Generative Adversarial Networkscited by 464
  2. 2014Theoretical guarantees for approximate sampling from smooth and log-concave densitiescited by 208
  3. 2015A Differential Equation for Modeling Nesterov's Accelerated Gradient Method: Theory and Insightscited by 206
  4. 2017Stein Variational Gradient Descent as Gradient Flowcited by 152
  5. 2017Underdamped Langevin MCMC: A non-asymptotic analysiscited by 133
  6. 2019Accelerated Information Gradient flowcited by 39
  7. 2023Taming Mode Collapse in Score Distillation for Text-to-3D Generationcited by 36
  8. 2025No Trick, No Treat: Pursuits and Challenges Towards Simulation-free Training of Neural Samplerscited by 28
  9. 2017Fractional Langevin Monte Carlo: Exploring L'evy Driven Stochastic Differential Equations for Markov Chain Monte Carlocited by 13
  10. 2023Accelerating optimization over the space of probability measurescited by 10
  11. 2024Shifted Composition III: Local Error Framework for KL Divergencecited by 7
  12. 2025Accelerated Stein Variational Gradient Flowcited by 6
7 more
  1. 2021Stein Variational Gradient Descent with Multiple Kernelcited by 4
  2. 2022A sharp uniform-in-time error estimate for Stochastic Gradient Langevin Dynamicscited by 4
  3. 2018Optimal Design of Process Flexibility for General Production Systemscited by 3
  4. 2020Preconditioned accelerated gradient descent methods for locally Lipschitz smooth objectives with applications to the solution of nonlinear PDEscited by 3
  5. 2025Diffusion-based supervised learning of generative models for efficient sampling of multimodal distributionscited by 2
  6. 2025Accelerated Markov Chain Monte Carlo Algorithms on Discrete Statescited by 2
  7. 2015Novel Bernstein-like Concentration Inequalities for the Missing Masscited by 1

Built on by

Not cited yet.