SA-PAL combines adaptive timesteps and position-dependent friction in Langevin dynamics, reporting 1.5-3x faster mixing on Rosenbrock and Mueller-Brown potentials plus order-of-magnitude efficiency gains on other test problems.
Monte Carlo Methods in Statistical Mechanics: Foundations and New Algorithms, pp.\ 131--192
3 Pith papers cite this work, alongside 545 external citations. Polarity classification is still indexing.
3
Pith papers citing it
545
external citations · Crossref
years
2026 3representative citing papers
Learned wavelet-based conditional sampling decorrelates within O(1) sweeps per scale at an Ising-like critical point of a frustrated 2D soft-spin model, giving O(log_2 L) total sampling complexity.
New discrete-time approximations to SG(L)D enable accurate non-asymptotic predictions of covariance and integrated autocorrelation time for practical tuning in large-batch or misspecified regimes.
citing papers explorer
-
Accelerated sampling using SamAdams variable timesteps and position-adaptive Langevin dynamics
SA-PAL combines adaptive timesteps and position-dependent friction in Langevin dynamics, reporting 1.5-3x faster mixing on Rosenbrock and Mueller-Brown potentials plus order-of-magnitude efficiency gains on other test problems.