[关闭]
@nrailgun 2016-08-25T15:26:15.000000Z 字数 574 阅读 1507

Auto-Encoder

机器学习


An auto encoder is an unsupervised learning algorithm that applies backpropagation, setting the target values to be equal to the inputs, i.e., y=x.

To impose a "sparsity" constraint on hidden units, we add a KL divergence term to the loss function.

js(ρlogρρ^j+(1ρ)log1ρ1ρ^j)

where ρ^j=1mmi[aj(x)], s is the number of hidden units.

Our overall loss function is

Jsparse(W,b)=J(W,b)+βjsKL(ρρ^j)

添加新批注
在作者公开此批注前,只有你和作者可见。
回复批注