Deep Bayes:MCMC
Introduction
这篇主要记录了Monto Carlo Markov Chain。转载请注明。
ref:
Deep Bayes slides
https://blog.csdn.net/qy20115549/article/details/54561643
MCMC
Probabilistic model:
MCMC methods construct samples from using only
Usage:
Markov chain properties
Gibbs sampling
- Can be applied both for discrete and continuous variables;
- Doesn't have any parameters for tuning
- Can be highly inefficient for large dimensional case
Metropolis-Hastings sampling
Hamiltonian equations
思想:
Summary
主要是学习了一下HMC。
算法小屋 文章被收录于专栏
不定期分享各类算法以及面经。同时也正在学习相关分布式技术。欢迎一起交流。