Adaptiveness of CMA based Samplers
| Autor: | Edna Chelangat Milgo, Peter Waiganjo, Nixon Ronoh, Bernard Manderick |
|---|---|
| Přispěvatelé: | Faculty of Sciences and Bioengineering Sciences, Informatics and Applied Informatics |
| Jazyk: | angličtina |
| Rok vydání: | 2017 |
| Předmět: |
Mathematical optimization
MCMC Sampling (statistics) Slice sampling Markov chain Monte Carlo 02 engineering and technology 01 natural sciences Statistics::Computation 010104 statistics & probability symbols.namesake Distribution (mathematics) Rate of convergence Mixing (mathematics) 0202 electrical engineering electronic engineering information engineering symbols Statistics::Methodology 020201 artificial intelligence & image processing 0101 mathematics CMA-ES Evolution strategy Mathematics Computer Science(all) |
| Zdroj: | GECCO (Companion) |
| Popis: | We turn the Covariance Matrix Adaptation Evolution Strategy into an adaptive Markov Chain Monte Carlo (or MCMC) sampling algorithm that adapts online to the target distribution, i.e. the distribution to be sampled from. We call the resulting algorithm CMA-Sampling. It exhibits a higher convergence rate, a better mixing, and consequently a more effective MCMC sampler. We look at a few variants and compare their adaptiveness to a number of other adaptive samplers, including Haario et. al's AM sampler, on a testsuite of 4 target distributions. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|