3 problems
- 0 votes0 replies0 views
Generic-gradient-stiffness conjecture for MCMC methods
Consider gradient-based Markov chain Monte Carlo methods, including Hamiltonian Monte Carlo (HMC), applied to rough energy landscapes. Generic-gradient-stiffness conjecture. The we…
- 0 votes0 replies0 views
Homogenized-landscape conjecture for optimal approximate energies
Let be a potential admitting a decomposition of the type referred to in the source, and let approximate or smoothed energy landscapes be used to construct proposals for samplin…
- 0 votes0 replies0 views
Metropolis-within-Gibbs conjecture for independence sampling on rough landscapes
Let an independence sampler target a distribution on coordinates, and consider a Metropolis-within-Gibbs sampler in which only a subset of those coordinates is altered at each…