贝叶斯神经网络 [英] Bayesian Neural Network

问题描述

先生，我不清楚贝叶斯神经网络.它是训练算法还是本身就是神经网络(如MLP或径向基函数网络)?
如果它是一个网络，那么其架构将如何.

预先感谢您的帮助.

Sangeeta

Sir, i am not clear about Bayesian Neural Network. Is it a training algorith or itself a Neural Network like MLP or Radial Basis Function Network?
If its a Network then how its architecture will be.

Thank you in advance for your help.

Sangeeta

推荐答案

请参阅 ^ ] Wikipedia的缓存副本可能会对您有所帮助.

See THIS[^] Cached copy of wikipedia may help you.

贝叶斯网络和贝叶斯神经网络是两种不同的事物.

通常，贝叶斯神经网络"是指在学习过程和正则化中在神经网络框架内使用贝叶斯方法.
它提供了一种不同的方式来解释训练过程，而不是使误差函数最小化，而是找到了最有可能描述数据的权重集.

可以找到误差条，并且可以基于证据框架比较不同的模型.

另一方面，贝叶斯网络与神经网络没有关系，它们是描述推理规则的图形模型.

Bayesian Networks and Bayesian Neural Networks are two different things.

Usually, "Bayesian Neural Networks" refers to the use of Bayesian methods within the Neural Network framework in the learning process and for regularization.

It provides a different way to interpret the training process, instead of minimizing an error function, you''re finding the most probably set of weights describing the data.

Error bars can be found and different models can be compared based on an evidence framework.

Bayesian Networks on the other hand bear no relationship to Neural Networks, they are Graphical Models describing inference rules.

这篇关于贝叶斯神经网络的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！