Three Samplers, One Posterior: NUTS vs. Emcee vs. Nested Sampling
中文摘要
本文通过与精确真值对比,评估了 NUTS、Emcee 和嵌套采样在估计后验分布时的表现及失效情况。
English Summary
This article benchmarks NUTS, Emcee, and Nested Sampling against exact ground truth, evaluating their performance and failures in estimating a posterior distribution.
Original Excerpt
Benchmarking NumPyro, affine-invariant MCMC, and dynesty against exact ground truth and watching them fail.