Samplers#
JESTER provides modern Bayesian sampling algorithms optimized for EOS inference with JAX acceleration. All samplers support GPU hardware and automatic differentiation.
Available Samplers#
- Sequential Monte Carlo (SMC)
Adaptive tempering with Random Walk or NUTS kernels. Recommended default.
- Nested Sampling (NS-AW)
Acceptance Walk variant for evidence computation and parameter estimation.
- FlowMC
Normalizing flow-enhanced MCMC for efficient exploration of complex posteriors.