PAWL-Forced Simulated Tempering
This is an incremental improvement for researchers in computational statistics or machine learning working on sampling methods.
The authors tackled the automation and improvement of simulated tempering algorithms by applying the parallel adaptive Wang-Landau (PAWL) algorithm, demonstrating through a simple example that it brings additional improvements via parallelization, adaptive proposals, and automated bin splitting.
In this short note, we show how the parallel adaptive Wang-Landau (PAWL) algorithm of Bornn et al. (2013) can be used to automate and improve simulated tempering algorithms. While Wang-Landau and other stochastic approximation methods have frequently been applied within the simulated tempering framework, this note demonstrates through a simple example the additional improvements brought about by parallelization, adaptive proposals and automated bin splitting.